[{"data":1,"prerenderedAt":1986},["ShallowReactive",2],{"footer-vacatures-en":3,"hub-vastgelopen-met-ai-en":4,"content-by-keys-offers-en-vibe-code-rescue-audit,rescue-sprint,build-integrate":86,"content-by-keys-useCases-en-vibe-coded-app-professionaliseren,ai-prototype-naar-productie,ai-agent-niet-betrouwbaar,gegenereerde-code-niet-meer-begrijpelijk,ai-code-breekt-steeds,vast-op-auth-database-of-rechten,ai-koppelen-aan-bestaande-systemen,security-en-privacy-ai-oplossing,ai-oplossing-te-langzaam,ai-oplossing-te-duur,tests-evals-monitoring-voor-ai,developmentteam-loopt-vast-met-ai,technische-second-opinion-ai,ai-project-overnemen-van-andere-leverancier,ai-experiment-heeft-echte-gebruikers,prototype-naar-onderhoudbaar-softwareproduct,weg-bij-ai-platform-of-leverancier,vibe-code-project-vastgelopen":473,"content-by-keys-buyerTriggers-en-starten-met-ai,ai-laten-bouwen":1845,"content-by-keys-cases-en-":1965,"why-9to5-en":1966},[],{"id":5,"title":6,"approach":7,"body":22,"cta":29,"description":26,"extension":32,"hero":33,"key":35,"label":36,"meta":37,"navigation":38,"path":39,"proof":40,"relations":52,"seo":77,"stem":80,"summary":81,"visuals":82,"__hash__":85},"buyerTriggers_en\u002Fbuyer-triggers\u002Fvastgelopen-met-ai.md","We've already built something, but we're stuck.",{"title":8,"steps":9},"We step in, understand, repair and hand over.",[10,13,16,19],{"title":11,"text":12},"Confirm the rescue goal","We establish what is actually going wrong, what is urgent and what needs to be demonstrably better after the rescue.\n",{"title":14,"text":15},"Understand the context and code","We examine the code, architecture, data flows, integrations and deployment until the source of the problem is clear.\n",{"title":17,"text":18},"Fix and test","We make targeted repairs, add tests, evals and guardrails where needed and check that the solution works under real conditions too.\n",{"title":20,"text":21},"Document and hand over","We document decisions, clear critical technical debt and make sure your team can move forward independently again.\n",{"type":23,"value":24,"toc":25},"minimark",[],{"title":26,"searchDepth":27,"depth":27,"links":28},"",2,[],{"title":30,"body":31},"Stuck with AI or vibe-coded software?","Briefly describe what you have and where it is going wrong. We'll take a look and quickly tell you what makes sense to repair, keep or rebuild.\n","md",{"intro":34},"You don't have to start over just because things have become complicated. We step in, including when you have a vibe-coded app or code generated by AI tools such as Claude Code, Codex or Cursor. We understand what is there, find what is really going wrong and make it technically manageable again.\n","vastgelopen-met-ai","Stuck with AI",{},true,"\u002Fbuyer-triggers\u002Fvastgelopen-met-ai",{"intro":41,"points":42},"Assessing an existing codebase takes more than an automated quality score. It takes senior engineers who can identify the problems that really hurt and make targeted repairs.\n",[43,46,49],{"title":44,"text":45},"We work with existing software","A messy or AI-generated codebase is no reason to automatically start from scratch.\n",{"title":47,"text":48},"Software engineering and AI","We look at everyday production problems and AI-specific failure modes, such as unreliable output and missing evals.\n",{"title":50,"text":51},"Move forward independently again","We document decisions and clear critical technical debt so your team can move forward on its own again.\n",{"useCaseKeys":53,"offerKeys":72,"caseKeys":76},[54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71],"vibe-coded-app-professionaliseren","ai-prototype-naar-productie","ai-agent-niet-betrouwbaar","gegenereerde-code-niet-meer-begrijpelijk","ai-code-breekt-steeds","vast-op-auth-database-of-rechten","ai-koppelen-aan-bestaande-systemen","security-en-privacy-ai-oplossing","ai-oplossing-te-langzaam","ai-oplossing-te-duur","tests-evals-monitoring-voor-ai","developmentteam-loopt-vast-met-ai","technische-second-opinion-ai","ai-project-overnemen-van-andere-leverancier","ai-experiment-heeft-echte-gebruikers","prototype-naar-onderhoudbaar-softwareproduct","weg-bij-ai-platform-of-leverancier","vibe-code-project-vastgelopen",[73,74,75],"vibe-code-rescue-audit","rescue-sprint","build-integrate",[],{"title":78,"description":79},"Stuck with AI? Rescue by senior engineers | 9to5 software","Stuck with an AI agent, prototype or vibe-coded app (Lovable, Cursor, Bolt, Replit)? We analyse the problem, make targeted repairs and restore maintainability.\n","buyer-triggers\u002Fvastgelopen-met-ai","Something is already there, but it is unreliable, hard to understand or technically stuck. We need to get it back under control.\n",{"hero":83,"proof":84,"approach":84},"sq-tangle-knot","ls-rescue-repair","e0bDu3r1P4JqQOadOsQgPsL_EDOsQGpgsqBXrdwFTNw",[87,215,331],{"id":88,"title":89,"body":90,"catalogue":94,"cta":95,"description":26,"extension":32,"fit":100,"hero":104,"key":73,"meta":107,"navigation":38,"next":108,"outcome":120,"path":132,"price":133,"process":136,"proof":150,"relations":163,"scope":168,"seo":201,"stem":204,"summary":205,"summaryPanel":206,"visuals":211,"__hash__":214},"offers_en\u002Foffers\u002Fvibe-code-rescue-audit.md","Vibe Code Rescue Audit",{"type":23,"value":91,"toc":92},[],{"title":26,"searchDepth":27,"depth":27,"links":93},[],null,{"title":96,"body":97,"primaryCta":98,"local":99},"No idea what to fix anymore?","Let us first figure out what you actually have. Then you can make a technical decision without blindly prompting on or throwing everything away.\n","Book a Rescue Audit","Vibe Code Rescue by senior engineers from Delft, for teams in Delft, The Hague, Rotterdam and the surrounding area.\n",{"title":101,"intro":102,"useCaseKeys":103},"When is this offer a fit?","When you sense something is technically wrong, but the actual cause or extent isn't clear yet.\n",[54,56,57,71],{"headline":105,"intro":106,"primaryCta":98},"Find out what is really wrong before you rebuild everything.","You already have software, but no longer know which problems are superficial and which are structural. We reconstruct how the system works, separate technical debt from real risks and give you a prioritised route forward.\n",{},{"title":109,"items":110},"What happens next?",[111,114,117],{"title":112,"text":113,"offerKey":74},"A clear rescue goal?","Then we can tackle it in a focused Rescue Sprint.\n",{"title":115,"text":116,"offerKey":75},"Turns out to be a bigger build project?","Then we scope Build & Integrate as a fixed engagement.\n",{"title":118,"text":119},"Turns out little is wrong?","Then we'll simply tell you. A follow-up is not required.\n",{"title":121,"items":122},"Afterwards you have...",[123,126,129],{"title":124,"text":125},"Technical grip on what is there","Architecture, key dependencies and critical risks are clear again.\n",{"title":127,"text":128},"A prioritised rescue plan","You know what can stay, what really needs to change and in which order.\n",{"title":130,"text":131},"A decision on the follow-up approach","Fixing, partially rebuilding, developing further at a larger scale or doing nothing becomes a conscious choice.\n","\u002Foffers\u002Fvibe-code-rescue-audit",{"label":134,"type":135},"€2,950 excl. VAT","fixed",{"title":137,"steps":138},"How it works.",[139,141,144,147],{"title":11,"text":140},"We determine which questions the Audit needs to answer and where the biggest uncertainty lies.\n",{"title":142,"text":143},"Reconstruct code + system","We read the architecture, code, deployment, data flows and relevant dependencies.\n",{"title":145,"text":146},"Prioritise problems","We distinguish cosmetic debt, development blockers and real production or security risks.\n",{"title":148,"text":149},"Rescue plan + debrief","You get a concrete route with what can stay, what needs to change and what makes sense next.\n",{"title":151,"intro":152,"points":153,"caseKeys":162},"Why should you believe this?","Assessing an existing codebase takes more than an automated quality score. It takes senior engineers who can see which problems actually hurt.\n",[154,157,159],{"title":155,"text":156},"Existing software is where we work","A messy or AI-generated codebase is no reason to automatically start over.\n",{"title":47,"text":158},"We look at both regular production-grade problems and AI-specific failure modes.\n",{"title":160,"text":161},"Follow-up work is not a requirement","If the best conclusion is that your current route is fine, the engagement can end there.\n",[],{"buyerTriggerKeys":164,"useCaseKeys":165,"relatedOfferKeys":166,"caseKeys":167},[35],[54,55,56,57,58,61,62,63,66,67,68,69,70,71],[74,75],[],{"title":169,"intro":170,"activities":171,"deliverables":183,"exclusions":194},"What is included in this offer?","A technical diagnosis of one existing application or codebase, aimed at decision-making.\n",{"intro":172,"items":173},"We look broadly enough to find structural causes, without turning it into a theoretical code review.\n",[174,175,176,177,178,179,180,181,182],"Examining architecture and code quality","Assessing dependencies and technical debt","Reviewing authentication and authorisation","Examining the data layer and relevant integrations","Assessing deployment and error handling","Reviewing tests, logging and monitoring","Listing security risks at a high level","Examining AI-specific failure modes","Assessing maintainability and lock-in",{"intro":184,"items":185},"You get a technical verdict you can actually base a follow-up on.\n",[186,187,188,189,190,191,192,193],"Architecture overview","Prioritised list of problems and risks","What can stay","What needs to be adjusted","What may need to be redone","Advice: fix \u002F partially rebuild \u002F rebuild","Proposed order and size of the follow-up","Technical debrief with a senior engineer",{"intro":195,"items":196},"The Audit delivers understanding and a plan. Major fixes don't quietly become part of the analysis.\n",[197,198,199,200],"Full repair of the codebase","Complete rewrite","Unbounded pentest or formal security audit","Product development beyond small diagnostic changes",{"title":202,"description":203},"Vibe Code Rescue Audit | Stuck with AI code | 9to5 software","Vibe-coded app (Lovable, Cursor, Bolt, Replit) or AI project stuck? 9to5 examines architecture, code, security, tests and deployment and gives you a rescue plan.\n","offers\u002Fvibe-code-rescue-audit","We examine what you have, what can stay as it is, where the real risks are and how you can move forward responsibly again.\n",{"qualifier":207,"result":208,"duration":209,"nextStep":210},"Fixed price · one existing application or codebase","Technical grip on what exists and a prioritised rescue plan","About two weeks","A Rescue Sprint for the most important problem, or continue yourself",{"hero":83,"proof":212,"process":213},"sq-orient-options","ls-route-converge","IWSlwlvSU82rOHwWTXtSCvLbfXjNUybu2L5bZ-96IJY",{"id":216,"title":217,"body":218,"catalogue":94,"cta":222,"description":26,"extension":32,"fit":227,"hero":230,"key":74,"meta":233,"navigation":38,"next":234,"outcome":245,"path":256,"price":257,"process":260,"proof":273,"relations":286,"scope":291,"seo":318,"stem":321,"summary":322,"summaryPanel":323,"visuals":328,"__hash__":330},"offers_en\u002Foffers\u002Frescue-sprint.md","Rescue Sprint",{"type":23,"value":219,"toc":220},[],{"title":26,"searchDepth":27,"depth":27,"links":221},[],{"title":223,"body":224,"primaryCta":225,"local":226},"Do you know what's broken?","Share the technical problem with us. If it's sharp enough, we don't need to sell you an extra audit first.\n","Plan a Rescue Sprint","Temporary senior engineering from Delft, available for technical AI and software problems in the region.\n",{"title":101,"intro":228,"useCaseKeys":229},"When the technical problem is sharp enough to fix right away.\n",[59,60,64,65],{"headline":231,"intro":232,"primaryCta":225},"One clear technical problem. One week of senior engineering.","You know where the problem is, but your team can't solve it or lacks exactly the specialist experience needed right now. We step in with focus, fix the hard part and make sure you can move forward again afterwards.\n",{},{"title":109,"items":235},[236,239,242],{"title":237,"text":238},"Done? Then we leave.","If you can continue independently, the Sprint has succeeded.\n",{"title":240,"text":241,"offerKey":74},"Another clear rescue goal?","Then you can use a next Sprint as a new fixed scope.\n",{"title":243,"text":244,"offerKey":75},"Turns out bigger than a Sprint?","Then we scope Build & Integrate. If the cause is unclear after all, the Rescue Audit comes first.\n",{"title":121,"items":246},[247,250,253],{"title":248,"text":249},"The concrete problem is tackled","The agreed technical blocker is solved or demonstrably moved forward.\n",{"title":251,"text":252},"The change is set up professionally","Where needed we add tests, evals, guardrails or production preconditions.\n",{"title":254,"text":255},"Your team can move forward again","Code, choices and follow-up are documented and technically handed over.\n","\u002Foffers\u002Frescue-sprint",{"label":258,"suffix":259,"type":135},"€5,950 excl. VAT","per Sprint",{"title":137,"steps":261},[262,264,267,270],{"title":11,"text":263},"We check whether the problem is sharp enough for a Sprint.\n",{"title":265,"text":266},"Understand context + code","We read exactly enough to intervene responsibly.\n",{"title":268,"text":269},"Solve + test","We implement the fix and add the quality layer it needs.\n",{"title":271,"text":272},"Document + hand over","We record choices and make sure your team can continue independently again.\n",{"title":151,"intro":274,"points":275,"caseKeys":285},"This is exactly the kind of work senior engineers are valuable for: quickly understanding an unfamiliar codebase, solving the hard part and leaving again.\n",[276,279,282],{"title":277,"text":278},"Senior AI + software engineers","We step in where it gets complicated and leave when it's done.\n",{"title":280,"text":281},"We work in what is already there","We don't expect you to migrate everything to our favourite stack first.\n",{"title":283,"text":284},"Yours as soon as it's done","We document, hand over and leave as soon as you can continue on your own.\n",[],{"buyerTriggerKeys":287,"useCaseKeys":288,"relatedOfferKeys":289,"caseKeys":290},[35],[59,60,64,65],[73,75],[],{"title":169,"intro":292,"activities":293,"deliverables":302,"exclusions":311},"One Sprint has one clear rescue goal. That boundary keeps it fast and honest.\n",{"intro":294,"items":295},"We do what is technically needed to tackle the agreed problem responsibly.\n",[296,297,298,299,300,301],"Understanding the relevant code and context","Reproducing and confirming the root cause","Implementing a targeted fix or technical improvement","Adding relevant tests, evals or checks","Adjusting integration, deployment or data where the goal requires it","Recording and handing over technical choices",{"intro":303,"items":304},"Working engineering in the existing codebase, not an advice document someone else still has to carry out.\n",[305,306,307,308,309,310],"Implemented technical change","Relevant tests or quality checks","Working code in the agreed environment","Documentation of key choices","Handover to your own team","Clarity about any remaining follow-up",{"intro":312,"items":313},"A Sprint is bounded and doesn't quietly turn into an open-ended project.\n",[314,315,316,317],"Multiple separate rescue goals at once","Unbounded refactor of the whole codebase","Major new product features beyond the rescue goal","Automatic extension or retainer",{"title":319,"description":320},"Rescue Sprint | Senior AI & software engineering | 9to5 software","One week of senior engineering for a technical blocker in AI, an AI agent or vibe-coded software. Repair, test, document and hand over, €5,950 excl. VAT.\n","offers\u002Frescue-sprint","One week of senior engineering focused on one clear technical rescue goal. No retainer and no endless engagement.\n",{"qualifier":324,"result":325,"duration":326,"nextStep":327},"Per Sprint · one clear rescue goal","The agreed technical blocker solved, tested and handed over","One week","You take it over, or another Sprint for a new goal",{"hero":83,"proof":329,"process":84},"sq-guardrails-check","RBo-3uCvL4DWueBHoZFGrqKFPfwCZ1QLbywOlnLG0NY",{"id":332,"title":333,"body":334,"catalogue":338,"cta":349,"description":26,"extension":32,"fit":354,"hero":361,"key":75,"meta":364,"navigation":38,"next":365,"outcome":377,"path":388,"price":389,"process":392,"proof":406,"relations":419,"scope":430,"seo":459,"stem":462,"summary":463,"summaryPanel":464,"visuals":469,"__hash__":472},"offers_en\u002Foffers\u002Fbuild-integrate.md","Build & Integrate",{"type":23,"value":335,"toc":336},[],{"title":26,"searchDepth":27,"depth":27,"links":337},[],{"audience":339,"promise":340,"deliverables":341,"note":347,"visual":348},"For organisations that know what needs to be built and want 9to5 to build and integrate it professionally.\n","Not an AI demo, but something that works with your real software, data, processes and people.\n",[342,343,344,345,346],"Architecture and production-ready implementation","Integrations with existing software and data","Permissions, error handling and human oversight","Tests, monitoring and deployment where relevant","Documentation and handover to your team","We define the scope together first, then agree a fixed price for the result.\n","ls-system-connect",{"title":350,"body":351,"primaryCta":352,"local":353},"Do you know what needs to be built?","Share the desired outcome and your existing technical context with us. If the scope is clear, we turn it into a fixed engagement.\n","Discuss what you need built","AI engineering and software implementation from Delft, for organisations in Delft, The Hague, Rotterdam and the surrounding area.\n",{"title":101,"intro":355,"useCaseKeys":356},"When the desired outcome is clear enough to scope a professional implementation.\n",[357,358,359,360],"ai-agent-laten-bouwen","interne-ai-assistent-laten-ontwikkelen","bedrijfsproces-automatiseren-met-ai","ai-toevoegen-aan-bestaande-software",{"headline":362,"intro":363,"primaryCta":352},"Not an AI demo, but something that works with your real software, data and people.","The problem and the desired direction are clear. Now it has to be built properly. We take on architecture, AI, integrations and production-grade engineering as one whole and hand it back over afterwards.\n",{},{"title":109,"items":366},[367,370,373],{"title":368,"text":369},"You take it over","That is what we work towards. We don't want you to keep needing us.\n",{"title":371,"text":372},"A new bounded question comes up later","Then we scope it again instead of automatically keeping capacity running.\n",{"title":374,"text":375,"offerKey":376},"Still too much technical uncertainty?","Then an AI Prototype Sprint or Reality Check before Build & Integrate may be wiser.\n","ai-prototype-sprint",{"title":121,"items":378},[379,382,385],{"title":380,"text":381},"Working software instead of a demo","The agreed functionality works end to end within the real technical environment.\n",{"title":383,"text":384},"Reliable integration with what is already there","Data, systems, permissions and human processes are part of the solution.\n",{"title":386,"text":387},"A product you can hand over","Architecture, code and key choices are documented well enough to continue independently.\n","\u002Foffers\u002Fbuild-integrate",{"label":390,"type":391},"Fixed price after scoping","scoped",{"title":137,"steps":393},[394,397,400,403],{"title":395,"text":396},"Scope + success criteria","We make exact enough what needs to work to agree on responsibility and price.\n",{"title":398,"text":399},"Architecture + integrations","We determine how AI, data, existing software, permissions and users come together.\n",{"title":401,"text":402},"Build + test","We implement the hard part and also test errors, edge cases and reliability.\n",{"title":404,"text":405},"Go live + hand over","We bring the solution to production in a controlled way, document it and hand it over.\n",{"title":151,"intro":407,"points":408,"caseKeys":418},"AI only becomes truly valuable when it works together with existing software, data and business processes. That is simply serious software engineering.\n",[409,412,415],{"title":410,"text":411},"The same people advise and build","Advice and execution sit with the same technical people.\n",{"title":413,"text":414},"Integrations are core work","We don't treat APIs, databases, auth and existing systems as an afterthought around the AI.\n",{"title":416,"text":417},"Production and handover are part of it","The work isn't done when the demo works, but when the solution is manageable.\n",[],{"buyerTriggerKeys":420,"useCaseKeys":422,"relatedOfferKeys":427,"caseKeys":429},[421,35],"ai-laten-bouwen",[357,358,359,360,423,424,425,426],"eigen-data-en-kennis-gebruiken-met-ai","documenten-en-mails-automatisch-verwerken","agentic-workflows-bestaande-software","tools-vervangen-door-maatwerkflow",[428,376,73,74],"ai-reality-check",[],{"title":169,"intro":431,"activities":432,"deliverables":443,"exclusions":452},"The scope is defined up front. Then you get a fixed price for the agreed result.\n",{"intro":433,"items":434},"The exact work differs per engagement, but we take responsibility for the complete technical result.\n",[435,436,437,438,439,440,441,442],"Making scope and success criteria concrete","Designing the architecture","Implementing AI functionality or an agent flow","Building integrations with existing software and data","Setting up authentication, authorisation and data flows properly","Tests, evals, monitoring and error handling where relevant","Deployment and production setup","Documentation and technical handover",{"intro":444,"items":445},"The deliverables are defined per scope and are part of the fixed price.\n",[446,447,448,449,450,451],"Working agreed functionality","Production-grade implementation","Required integrations","Relevant tests and quality checks","Technical documentation","Deployment and handover as scoped",{"intro":453,"items":454},"No unbounded consultancy and no feature list that keeps growing afterwards.\n",[455,456,457,458],"Functionality outside the agreed scope","Ongoing retainer as a standard component","External model, API and hosting costs","Additional work without a new explicit scope and price",{"title":460,"description":461},"Have AI built and integrated | Build & Integrate | 9to5 software","Have an AI agent, internal tool, MCP integration or AI feature built with Claude or ChatGPT\u002FOpenAI and professionally integrated with your software, data and processes.\n","offers\u002Fbuild-integrate","You know what needs to be built. We build it professionally, connect it to your existing software and make sure it can be handed over production-ready.\n",{"qualifier":465,"result":466,"duration":467,"nextStep":468},"Fixed price after joint scoping","Working, integrated software your team can take over","Set in the scope","Handover to your team, or further development together",{"hero":470,"proof":471,"process":348},"sq-connect-nodes","sq-foundation-build","jlXsSL_a-hghGaOxvEHbZ2ARDUkoaAlfXaEkKLzQpSQ",[474,549,626,703,779,855,931,1007,1083,1161,1237,1314,1390,1466,1542,1618,1694,1769],{"id":475,"title":476,"body":477,"description":26,"desiredOutcome":481,"extension":32,"hero":493,"key":54,"meta":495,"navigation":38,"path":496,"problem":497,"proof":509,"recognition":520,"recommendedApproach":535,"relations":537,"seo":542,"stem":545,"summary":546,"visuals":547,"__hash__":548},"useCases_en\u002Fuse-cases\u002Fvibe-coded-app-professionaliseren.md","Our vibe-coded app needs to become professional",{"type":23,"value":478,"toc":479},[],{"title":26,"searchDepth":27,"depth":27,"links":480},[],{"intro":482,"items":483},"You want to keep the speed of AI-assisted development without your product becoming more and more fragile technically.\n",[484,487,490],{"title":485,"text":486},"A reliable technical foundation","Architecture, data, authentication, deployment and error handling are set up sensibly enough for serious use.\n",{"title":488,"text":489},"Able to keep changing safely","New features no longer unexpectedly break existing functionality every time.\n",{"title":491,"text":492},"Able to keep building yourself","You or your team understand the most important technical choices and can then carry on independently with AI tools again.\n",{"intro":494},"With Claude Code, Codex, Cursor, Lovable, Bolt or another AI tool you can get very far very fast. But as soon as real users, sensitive data or paying customers come in, technical choices suddenly matter. We help you determine what is fine, where the risks are and what is needed to move forward professionally.\n",{},"\u002Fuse-cases\u002Fvibe-coded-app-professionaliseren",{"intro":498,"causes":499},"Vibe coding isn't the problem. Treating a prototype as if it will become a production platform by itself usually is.\n",[500,503,506],{"title":501,"text":502},"The software grew faster than the architecture","Local solutions work individually, but together they no longer always form a logical system.\n",{"title":504,"text":505},"Technical guardrails are missing","Tests, clear boundaries, security checks and deployment feedback were often added later, or not at all.\n",{"title":507,"text":508},"The requirements have changed","What was fine for an experiment isn't automatically suitable for real users, sensitive data and continuous development.\n",{"intro":510,"points":511},"A code quality score won't get you far here. You need senior engineers who can understand existing software, prioritise risks and fix the hard part themselves.\n",[512,514,517],{"title":47,"text":513},"Besides the generated code, we look at the complete product around it.\n",{"title":515,"text":516},"We work in what's already there","Existing codebases, integrations and technical debt are no reason for us to start over.\n",{"title":518,"text":519},"Handover is part of the result","The goal is that after the rescue you can carry on independently and responsibly again.\n",{"intro":521,"signals":522},"The tipping point usually comes when an experiment slowly becomes a real software product.\n",[523,526,529,532],{"title":524,"text":525},"New changes feel riskier and riskier","A small feature can suddenly affect authentication, the database, deployment or other parts of the app.\n",{"title":527,"text":528},"The codebase gets harder to understand","There are more and more files, dependencies and solutions and nobody knows exactly why they were built that way.\n",{"title":530,"text":531},"Production is getting nerve-racking","You doubt whether data, permissions, backups, error handling and deployment are reliable enough for real users.\n",{"title":533,"text":534},"You want to move on, but not blindly","Rebuilding everything feels excessive, but carrying on as now no longer feels responsible either.\n",{"intro":536},"We start with a Vibe Code Rescue Audit. We examine what you have, what can stay as it is, where the real risks are and which improvements make the biggest difference. That way you avoid both building on blindly and needlessly redoing everything.\n",{"buyerTriggerKeys":538,"primaryOfferKey":73,"offerKeys":539,"caseKeys":540,"relatedUseCaseKeys":541},[35],[73,74],[],[71,69],{"title":543,"description":544},"Professionalise a vibe-coded app | 9to5 software","Ready to professionalise a vibe-coded app (Lovable, Bolt, Replit, v0)? 9to5 examines architecture, code, security, tests and deployment and turns it into serious software.\n","use-cases\u002Fvibe-coded-app-professionaliseren","You've built a surprising amount with AI tools. Now the app has to become reliable, secure and maintainable enough for serious use.\n",{"hero":83,"problem":471,"proof":84},"ESp-gjnPt-vPAQ3d4-MEUI8SZI--nHIB2va7yrp8svY",{"id":550,"title":551,"body":552,"description":26,"desiredOutcome":556,"extension":32,"hero":568,"key":55,"meta":570,"navigation":38,"path":571,"problem":572,"proof":584,"recognition":596,"recommendedApproach":611,"relations":613,"seo":618,"stem":621,"summary":622,"visuals":623,"__hash__":625},"useCases_en\u002Fuse-cases\u002Fai-prototype-naar-productie.md","Our AI prototype has to go to production",{"type":23,"value":553,"toc":554},[],{"title":26,"searchDepth":27,"depth":27,"links":555},[],{"intro":557,"items":558},"You want to keep what the prototype has proven, but reduce the technical risks before the system is used seriously.\n",[559,562,565],{"title":560,"text":561},"Reliable under real conditions","The solution can deal with errors, unexpected input and external systems that don't always cooperate.\n",{"title":563,"text":564},"Measurable behaviour","You can track whether quality, latency, costs and error rates stay within acceptable limits.\n",{"title":566,"text":567},"Manageable in production","Deployment, logging, security and human oversight are part of the system instead of separate aftercare.\n",{"intro":569},"A demo may act strangely now and then, production software may not. As soon as an AI prototype touches real users, data and business processes, the requirements change. We help determine what still lies between a successful experiment and a reliable production solution.\n",{},"\u002Fuse-cases\u002Fai-prototype-naar-productie",{"intro":573,"causes":574},"A prototype is built to prove an assumption. Production software has to deal with uncertainty, errors and real users for a long time.\n",[575,578,581],{"title":576,"text":577},"Reliability was not yet a main goal","During an experiment, speed of learning is usually more important than complete error handling.\n",{"title":579,"text":580},"AI does not behave deterministically","Good behaviour has to be measured and bounded instead of only being judged during a demo.\n",{"title":582,"text":583},"The rest of the software now counts","Data, permissions, integrations, deployment and monitoring become just as important as the AI functionality itself.\n",{"intro":585,"points":586},"For us, taking AI to production is mainly software engineering around a non-deterministic component. That calls for regular production-grade principles and AI-specific evaluation.\n",[587,590,593],{"title":588,"text":589},"Beyond the demo","We look at errors, edge cases, integrations and operational behaviour.\n",{"title":591,"text":592},"Human-in-the-loop where needed","Not every risk is solved with more AI.\n",{"title":594,"text":595},"Built for handover","Architecture and choices are recorded so that your own team can continue afterwards.\n",{"intro":597,"signals":598},"The prototype does what it should, but you notice there is still a big gap between demonstrating and going live responsibly.\n",[599,602,605,608],{"title":600,"text":601},"The happy flow works","Demos go well, but edge cases and errors have hardly been worked out.\n",{"title":603,"text":604},"You don't know what good enough is","There are no clear quality thresholds, evals or acceptance criteria yet.\n",{"title":606,"text":607},"Production questions pile up","Security, monitoring, costs, retries, permissions and support suddenly become important.\n",{"title":609,"text":610},"The first real users are coming","You want to prevent them from discovering the technical weak spots for you.\n",{"intro":612},"With a Vibe Code Rescue Audit we first map what stands between the prototype and production. If the route is then large and clear, Build & Integrate can deliver the production version in a controlled way.\n",{"buyerTriggerKeys":614,"primaryOfferKey":73,"offerKeys":615,"caseKeys":616,"relatedUseCaseKeys":617},[35],[73,75],[],[68,64],{"title":619,"description":620},"Take an AI prototype to production | 9to5 software","From a working AI prototype to production-ready software. 9to5 helps with architecture, reliability, security, integrations, monitoring and deployment.\n","use-cases\u002Fai-prototype-naar-productie","The prototype proves the idea works. Now it has to become reliable, secure and manageable enough for real users and real business processes.\n",{"hero":471,"problem":329,"proof":624},"ls-prototype-to-production","jSQ8HzLPrDfgmJ9vRmp8QviUxczEzNXGhTdx2Q7mSaU",{"id":627,"title":628,"body":629,"description":26,"desiredOutcome":633,"extension":32,"hero":645,"key":56,"meta":647,"navigation":38,"path":648,"problem":649,"proof":661,"recognition":673,"recommendedApproach":688,"relations":690,"seo":695,"stem":698,"summary":699,"visuals":700,"__hash__":702},"useCases_en\u002Fuse-cases\u002Fai-agent-niet-betrouwbaar.md","Our agent is not reliable enough",{"type":23,"value":630,"toc":631},[],{"title":26,"searchDepth":27,"depth":27,"links":632},[],{"intro":634,"items":635},"No agent is perfect all the time. You do need to know where it can fail and how the system deals with that.\n",[636,639,642],{"title":637,"text":638},"More predictable behaviour","The agent more consistently chooses the right steps, tools and information.\n",{"title":640,"text":641},"Errors become visible","You can measure which situations go wrong instead of relying on scattered manual tests.\n",{"title":643,"text":644},"Risks are contained","Validation, guardrails and human oversight prevent a mistake from immediately becoming a big problem.\n",{"intro":646},"An agent that works well eight out of ten times can look fantastic in a demo and be unusable in production. We find out where the unreliability comes from and which technical measures are needed to make the system manageable.\n",{},"\u002Fuse-cases\u002Fai-agent-niet-betrouwbaar",{"intro":650,"causes":651},"The cause rarely sits in the prompt alone. An agent is a complete software system in which model, tools, context, state and error handling work together.\n",[652,655,658],{"title":653,"text":654},"The task is too open","The agent gets more freedom than it needs to carry out the process well.\n",{"title":656,"text":657},"The feedback loop is missing","Without evals and traces it is hard to see which part is actually failing.\n",{"title":659,"text":660},"Errors are not caught","A single wrong model decision can needlessly carry through into every following step.\n",{"intro":662,"points":663},"You don't get a reliable agent by endlessly tweaking a single prompt. You treat it as software, with clear responsibilities, observability and controlled failure paths.\n",[664,667,670],{"title":665,"text":666},"Looking at the whole system","We examine the entire agent flow, not just the instruction text.\n",{"title":668,"text":669},"Measure first, then optimise","Evals and traces show which changes actually help.\n",{"title":671,"text":672},"Autonomy doesn't have to be the goal","We deliberately choose where deterministic software or human oversight is the wiser option.\n",{"intro":674,"signals":675},"At first it often feels like a series of separate incidents.\n",[676,679,682,685],{"title":677,"text":678},"The same input gives different results","What went well yesterday can suddenly take a different route today.\n",{"title":680,"text":681},"The agent sometimes picks the wrong tool","It calls actions that are not needed or uses the wrong data.\n",{"title":683,"text":684},"Prompt changes move the problem around","An improvement for one situation makes another situation worse.\n",{"title":686,"text":687},"Errors mostly show up with real users","The demos are fine, but reality contains far more variation.\n",{"intro":689},"In a Vibe Code Rescue Audit we first make the behaviour understandable and measurable. Then we determine which combination of architecture, evals, validation, guardrails and human-in-the-loop is needed. The concrete fixes can then follow in a Rescue Sprint.\n",{"buyerTriggerKeys":691,"primaryOfferKey":73,"offerKeys":692,"caseKeys":693,"relatedUseCaseKeys":694},[35],[73,74],[],[64,62],{"title":696,"description":697},"AI agent not reliable? | 9to5 software","Does your AI agent sometimes work well and sometimes not? 9to5 examines tools, data flows, evals, guardrails and architecture and makes behaviour more manageable.\n","use-cases\u002Fai-agent-niet-betrouwbaar","The agent can do impressive things, but does not behave predictably enough to trust a serious process or product to it.\n",{"hero":701,"problem":329,"proof":84},"sq-process-steps","msXea1UQcF1EwGJ9TUI7fGQMHikENHvoI9katWwH8e0",{"id":704,"title":705,"body":706,"description":26,"desiredOutcome":710,"extension":32,"hero":722,"key":57,"meta":724,"navigation":38,"path":725,"problem":726,"proof":738,"recognition":750,"recommendedApproach":765,"relations":767,"seo":772,"stem":775,"summary":776,"visuals":777,"__hash__":778},"useCases_en\u002Fuse-cases\u002Fgegenereerde-code-niet-meer-begrijpelijk.md","Nobody understands the generated code anymore",{"type":23,"value":707,"toc":708},[],{"title":26,"searchDepth":27,"depth":27,"links":709},[],{"intro":711,"items":712},"You don't need to know every file by heart. But an engineer does need to be able to understand where functionality belongs and what a change affects.\n",[713,716,719],{"title":714,"text":715},"An understandable structure","Responsibilities and boundaries in the codebase are recognisable again.\n",{"title":717,"text":718},"Fewer surprises","Engineers can better predict what a change affects elsewhere.\n",{"title":720,"text":721},"AI tools become useful again","Good structure also gives coding agents better context to work with responsibly.\n",{"intro":723},"AI can produce code faster than a team can understand it. That works until every change first requires archaeological research. We help you bring the codebase back to software people can get a grip on again.\n",{},"\u002Fuse-cases\u002Fgegenereerde-code-niet-meer-begrijpelijk",{"intro":727,"causes":728},"Code that compiles is not automatically a coherent software system.\n",[729,732,735],{"title":730,"text":731},"Every task was solved locally","Today's solution doesn't always fit last month's choices.\n",{"title":733,"text":734},"Architecture stayed implicit","Technical boundaries only exist in developers' heads, or not at all.\n",{"title":736,"text":737},"Cleaning up kept being postponed","AI makes adding cheap, so removing and consolidating get relatively little attention.\n",{"intro":739,"points":740},"With a messy codebase, the main job is judging what is essential, which debt really hurts, and how to stop the clean-up from becoming the new problem.\n",[741,744,747],{"title":742,"text":743},"Understand first","We don't blindly change code before it's clear what role that code plays.\n",{"title":745,"text":746},"Risk over beauty","Not every ugly construction needs to be rewritten right away.\n",{"title":748,"text":749},"Document what matters","Important architecture choices are made explicit so knowledge doesn't disappear again.\n",{"intro":751,"signals":752},"The code is usually readable enough. It's just that nobody understands how the whole fits together anymore.\n",[753,756,759,762],{"title":754,"text":755},"There are several solutions to the same problem","Old and new implementations live side by side and nobody knows which one leads.\n",{"title":757,"text":758},"Files and components keep growing","More and more responsibilities end up in the same places.\n",{"title":760,"text":761},"Deleting feels dangerous","Nobody knows for sure whether seemingly unused code is needed somewhere after all.\n",{"title":763,"text":764},"Every prompt needs enormous context","You have to explain to the AI over and over how large parts of the project work.\n",{"intro":766},"A Vibe Code Rescue Audit first maps the actual structure, duplication and technical risks. Then we only tackle the parts that actually block development. We don't refactor for the sake of it.\n",{"buyerTriggerKeys":768,"primaryOfferKey":73,"offerKeys":769,"caseKeys":770,"relatedUseCaseKeys":771},[35],[73,74],[],[58,71],{"title":773,"description":774},"Generated code no longer understandable? | 9to5 software","AI-generated code that nobody understands any more? 9to5 maps the architecture and technical debt and makes the codebase manageable again.\n","use-cases\u002Fgegenereerde-code-niet-meer-begrijpelijk","The software grows fast, but its structure and technical choices have become so opaque that changes are getting riskier and riskier.\n",{"hero":83,"problem":471,"proof":84},"Xz-f4kDEFuXwEdtsDZ6LRB0w0C3Rm0d5OcJI8Pk0jts",{"id":780,"title":781,"body":782,"description":26,"desiredOutcome":786,"extension":32,"hero":798,"key":58,"meta":800,"navigation":38,"path":801,"problem":802,"proof":814,"recognition":826,"recommendedApproach":841,"relations":843,"seo":848,"stem":851,"summary":852,"visuals":853,"__hash__":854},"useCases_en\u002Fuse-cases\u002Fai-code-breekt-steeds.md","Every new AI code change breaks something else",{"type":23,"value":783,"toc":784},[],{"title":26,"searchDepth":27,"depth":27,"links":785},[],{"intro":787,"items":788},"You want to be able to make changes again without discovering afterwards, every time, what collateral damage they caused.\n",[789,792,795],{"title":790,"text":791},"Changes stay local","A change only has effects in the places where you expect them.\n",{"title":793,"text":794},"Regressions are found quickly","Tests and checks flag problems before users discover them.\n",{"title":796,"text":797},"Development becomes truly fast again","Less repair work means AI-assisted development saves time once more.\n",{"intro":799},"You ask for one small change and three other things stop working. That is usually not bad luck. It points to missing boundaries, tests or technical coherence. We find out where that fragility comes from.\n",{},"\u002Fuse-cases\u002Fai-code-breekt-steeds",{"intro":803,"causes":804},"AI coding tools are strong at local changes. They don't automatically know all the implicit agreements that run through a growing product.\n",[805,808,811],{"title":806,"text":807},"Tests don't cover critical agreements","The AI gets little feedback when a change breaks existing functionality.\n",{"title":809,"text":810},"Parts are too tightly coupled","A local change affects state, data or behaviour in unexpected places.\n",{"title":812,"text":813},"Interfaces are not clear enough","The codebase doesn't show well enough which contracts must stay stable.\n",{"intro":815,"points":816},"The AI doesn't need to write less code. The codebase needs enough feedback and structure to make errors visible quickly.\n",[817,820,823],{"title":818,"text":819},"Tests where they add value","We test critical contracts and flows, not every detail for a percentage.\n",{"title":821,"text":822},"Clear technical boundaries","Fewer hidden couplings means fewer unexpected regressions.\n",{"title":824,"text":825},"Practical AI guardrails","The development workflow helps coding agents stay within the architecture.\n",{"intro":827,"signals":828},"The pattern often becomes visible once the codebase grows and several parts depend on each other.\n",[829,832,835,838],{"title":830,"text":831},"A simple feature causes unexpected bugs","Functionality on other screens or flows suddenly stops working.\n",{"title":833,"text":834},"The same bug comes back later","A bug gets fixed, but nothing prevents it from reappearing.\n",{"title":836,"text":837},"Refactors are almost impossible","Too many parts turn out to depend on each other implicitly.\n",{"title":839,"text":840},"You mostly rely on clicking through by hand","After every change someone has to walk through the whole application to see whether everything still works.\n",{"intro":842},"A Vibe Code Rescue Audit first finds out why changes cause so much collateral damage. We can then tackle the most important technical causes in a focused Rescue Sprint.\n",{"buyerTriggerKeys":844,"primaryOfferKey":73,"offerKeys":845,"caseKeys":846,"relatedUseCaseKeys":847},[35],[73,74],[],[57,64],{"title":849,"description":850},"AI code keeps breaking something else | 9to5 software","Does AI-generated code keep breaking existing functionality? 9to5 examines architecture, tests and dependencies and makes changes safer.\n","use-cases\u002Fai-code-breekt-steeds","New features ship quickly, but regressions, unexpected errors and repair work take more and more time.\n",{"hero":83,"problem":329,"proof":84},"Wac_TNYZVCtWw1yfVa0Y41CtfZ--5euF9AMy8eplg3g",{"id":856,"title":857,"body":858,"description":26,"desiredOutcome":862,"extension":32,"hero":874,"key":59,"meta":876,"navigation":38,"path":877,"problem":878,"proof":890,"recognition":902,"recommendedApproach":917,"relations":919,"seo":924,"stem":927,"summary":928,"visuals":929,"__hash__":930},"useCases_en\u002Fuse-cases\u002Fvast-op-auth-database-of-rechten.md","We're stuck on authentication, the database or permissions",{"type":23,"value":859,"toc":860},[],{"title":26,"searchDepth":27,"depth":27,"links":861},[],{"intro":863,"items":864},"You don't want a complete redesign. You want the concrete technical blocker to disappear without introducing new problems.\n",[865,868,871],{"title":866,"text":867},"The blocker is solved","The auth, database or permission flow in question demonstrably works as intended.\n",{"title":869,"text":870},"The solution is safe","We fix the problem without compromising data integrity or access.\n",{"title":872,"text":873},"Your team can move on","The relevant choices and points of attention are clear enough to carry on developing yourselves.\n",{"intro":875},"Sometimes the problem isn't vague at all. Login doesn't work properly, permissions leak, a migration fails or your data model gets in the way. Then you don't need a big advisory engagement, but someone who solves the hard technical part.\n",{},"\u002Fuse-cases\u002Fvast-op-auth-database-of-rechten",{"intro":879,"causes":880},"Authentication, authorisation and databases are fundamental parts of software. A quick workaround can therefore cause much bigger problems.\n",[881,884,887],{"title":882,"text":883},"The problem touches several layers","Frontend, backend, provider, database and permission model all need the same picture of the user.\n",{"title":885,"text":886},"Mistakes have serious consequences","A wrong fix can lead to data loss or access to data that should not be visible.\n",{"title":888,"text":889},"External providers add complexity","Auth and database platforms abstract a lot away, until you leave their happy flow.\n",{"intro":891,"points":892},"These are ordinary senior software engineering problems. That's exactly why they don't necessarily need a whole AI engagement around them.\n",[893,896,899],{"title":894,"text":895},"Straight into the code","No weeks of analysis when the technical problem is already sharp enough.\n",{"title":897,"text":898},"Security counts","A working login isn't enough if the permissions are wrong afterwards.\n",{"title":900,"text":901},"No permanent dependency","We fix, test, document and hand over again.\n",{"intro":903,"signals":904},"This is usually a well-defined technical problem that can still eat up a surprising amount of time.\n",[905,908,911,914],{"title":906,"text":907},"Login or OAuth doesn't work reliably","Redirects, callbacks, tokens or sessions behave differently than expected.\n",{"title":909,"text":910},"Roles and permissions are unclear","Users can do too much, too little or inconsistently different things.\n",{"title":912,"text":913},"The data model gets in the way","New functionality doesn't fit well with the existing tables and relations.\n",{"title":915,"text":916},"Migrations or data problems block deployment","You can't safely continue without first solving the underlying database problems.\n",{"intro":918},"If the problem is clear enough during the intake, we start a Rescue Sprint straight away. We reproduce the problem, understand the relevant code and systems and fix the blocker in a targeted way. Only if the cause turns out to be much more fundamental is a broader audit needed first.\n",{"buyerTriggerKeys":920,"primaryOfferKey":74,"offerKeys":921,"caseKeys":922,"relatedUseCaseKeys":923},[35],[74,73],[],[61,58],{"title":925,"description":926},"Stuck on authentication, database or permissions? | 9to5 software","Stuck on auth, the database or authorisation? Senior engineers from 9to5 step in and solve the technical blocker.\n","use-cases\u002Fvast-op-auth-database-of-rechten","The feature is clear, but a concrete technical problem around login, data or authorisation is holding the project back.\n",{"hero":83,"problem":329,"proof":84},"xePc9f5VpI4uja65DBv7Mo4zT3X8M-utSIVOkv5ZLHs",{"id":932,"title":933,"body":934,"description":26,"desiredOutcome":938,"extension":32,"hero":950,"key":60,"meta":952,"navigation":38,"path":953,"problem":954,"proof":966,"recognition":978,"recommendedApproach":993,"relations":995,"seo":1000,"stem":1003,"summary":1004,"visuals":1005,"__hash__":1006},"useCases_en\u002Fuse-cases\u002Fai-koppelen-aan-bestaande-systemen.md","We can't get the solution connected to existing systems",{"type":23,"value":935,"toc":936},[],{"title":26,"searchDepth":27,"depth":27,"links":937},[],{"intro":939,"items":940},"The AI shouldn't stay an isolated island. It should become a reliable part of the systems and processes that already exist.\n",[941,944,947],{"title":942,"text":943},"Reliable data exchange","Data arrives at the right moment, in the right format and at the right system.\n",{"title":945,"text":946},"Errors are manageable","Timeouts, duplicates and temporary outages don't immediately lead to unpredictable behaviour.\n",{"title":948,"text":949},"Permissions stay correct","The integration respects existing users, roles and access rules.\n",{"intro":951},"The AI demo works. It just still needs to fetch data from system A, change something in system B and respect permissions from system C. That is where an AI project often becomes a serious software engineering project.\n",{},"\u002Fuse-cases\u002Fai-koppelen-aan-bestaande-systemen",{"intro":955,"causes":956},"An integration is more than a line between two boxes in an architecture diagram. It also has to cope with real-world conditions.\n",[957,960,963],{"title":958,"text":959},"Systems speak different languages","Data models, statuses and responsibilities rarely map one-to-one.\n",{"title":961,"text":962},"External systems are not always available","Retries, queues and error handling become important as soon as the integration reaches production.\n",{"title":964,"text":965},"AI actions have real consequences","When an agent is allowed to change something, it must be clear who is responsible and what happens when something goes wrong.\n",{"intro":967,"points":968},"Building integrations is at the core of how 9to5 software works. AI only becomes useful when it can work safely with real data, software and people.\n",[969,972,975],{"title":970,"text":971},"Existing software is the starting point","The organisation doesn't have to adapt to the AI solution first.\n",{"title":973,"text":974},"Engineering around the API","We also design retries, error handling, logging and permissions.\n",{"title":976,"text":977},"No demo integrations","The integration has to stay usable under real conditions.\n",{"intro":979,"signals":980},"Often the AI part has already been proven and the real technical blocker lies in everything around it.\n",[981,984,987,990],{"title":982,"text":983},"The API doesn't quite do what you need","Documentation is limited or outdated, or the desired flow doesn't map neatly onto the available endpoints.\n",{"title":985,"text":986},"Authentication between systems gets stuck","Tokens, service accounts or user delegation make the integration more complicated.\n",{"title":988,"text":989},"Data drifts apart","Different systems use different identifiers, statuses or data models.\n",{"title":991,"text":992},"The happy flow works, errors don't","As soon as an external system is temporarily unavailable, the workflow loses track.\n",{"intro":994},"If the desired integration is clear, we tackle it directly in a Rescue Sprint. If the intake shows that a completely new integrated component is actually needed, Build & Integrate is a better fit.\n",{"buyerTriggerKeys":996,"primaryOfferKey":74,"offerKeys":997,"caseKeys":998,"relatedUseCaseKeys":999},[35],[74,75],[],[59,67],{"title":1001,"description":1002},"Connect ChatGPT or Claude to your software | 9to5 software","Connect ChatGPT, Claude or OpenAI to your software? 9to5 connects AI to existing APIs, databases and business processes.\n","use-cases\u002Fai-koppelen-aan-bestaande-systemen","The AI works on its own, but the real value sits behind APIs, databases and business software where the integration gets stuck.\n",{"hero":470,"problem":83,"proof":348},"4sCin98vwNyw5GXs3xalBs9ncmpDvENKXA0rH4xhya4",{"id":1008,"title":1009,"body":1010,"description":26,"desiredOutcome":1014,"extension":32,"hero":1026,"key":61,"meta":1028,"navigation":38,"path":1029,"problem":1030,"proof":1042,"recognition":1054,"recommendedApproach":1069,"relations":1071,"seo":1076,"stem":1079,"summary":1080,"visuals":1081,"__hash__":1082},"useCases_en\u002Fuse-cases\u002Fsecurity-en-privacy-ai-oplossing.md","We're worried about security or privacy",{"type":23,"value":1011,"toc":1012},[],{"title":26,"searchDepth":27,"depth":27,"links":1013},[],{"intro":1015,"items":1016},"You want to be able to explain technically where data goes and who may do what, so you don't have to hope sensitive data is handled well.\n",[1017,1020,1023],{"title":1018,"text":1019},"Clear data flows","You know which data stays internal and which external services process it.\n",{"title":1021,"text":1022},"Control over access","Authentication, authorisation and tenant isolation are set up deliberately.\n",{"title":1024,"text":1025},"Concrete risk priorities","You know which problems are truly urgent and which risks are acceptable.\n",{"intro":1027},"Which data are we actually sending to which model? Can users see each other's data? What do we log? Where are the secrets? If you can't answer those questions convincingly, it's time to take a proper look at the technical situation.\n",{},"\u002Fuse-cases\u002Fsecurity-en-privacy-ai-oplossing",{"intro":1031,"causes":1032},"AI adds new data flows and external dependencies to ordinary software problems around security and privacy.\n",[1033,1036,1039],{"title":1034,"text":1035},"Data moves through more systems","Model providers, vector stores, observability tools and integrations can all be part of the same flow.\n",{"title":1037,"text":1038},"AI input is less predictable","Users can provide unexpected content or try to influence system behaviour.\n",{"title":1040,"text":1041},"The underlying software remains decisive","Poor authorisation isn't fixed by choosing a safer language model.\n",{"intro":1043,"points":1044},"For us, AI security starts with ordinary software engineering, not with prompts: understanding data flows, limiting access, managing secrets and containing errors.\n",[1045,1048,1051],{"title":1046,"text":1047},"End-to-end view","From user and database to model provider and logging.\n",{"title":1049,"text":1050},"Practical prioritisation","Not every theoretical risk gets the same urgency.\n",{"title":1052,"text":1053},"Able to build the fix too","The advice doesn't have to go to another technical party afterwards.\n",{"intro":1055,"signals":1056},"Security questions often only come up seriously when an experiment moves towards production or sensitive data.\n",[1057,1060,1063,1066],{"title":1058,"text":1059},"Nobody knows exactly which data goes to the model","Prompts, context and uploads are assembled in several places in the application.\n",{"title":1061,"text":1062},"Roles and permissions feel fragile","It's hard to demonstrate that users can only reach their own data.\n",{"title":1064,"text":1065},"Secrets and external services grew organically","API keys and vendors were added whenever they were needed.\n",{"title":1067,"text":1068},"Logging may contain sensitive content","Debugging information may expose more than is desirable.\n",{"intro":1070},"A Vibe Code Rescue Audit maps the technical risks, data flows, permissions and external dependencies. Concrete improvements can then be carried out in a focused Rescue Sprint.\n",{"buyerTriggerKeys":1072,"primaryOfferKey":73,"offerKeys":1073,"caseKeys":1074,"relatedUseCaseKeys":1075},[35],[73,74],[],[59,70],{"title":1077,"description":1078},"Assess the security and privacy of an AI solution | 9to5 software","Concerned about security, privacy or data flows in your AI solution? 9to5 maps technical risks, permissions, vendors and data flows.\n","use-cases\u002Fsecurity-en-privacy-ai-oplossing","The solution works, but it isn't clear enough where data goes, who can access it and what technical risks there are.\n",{"hero":329,"problem":83,"proof":84},"kwq5CRw0CzrlXbFE7Foamr17AqEFniwI0gkZ6zHai1A",{"id":1084,"title":1085,"body":1086,"description":26,"desiredOutcome":1090,"extension":32,"hero":1102,"key":62,"meta":1104,"navigation":38,"path":1105,"problem":1106,"proof":1118,"recognition":1130,"recommendedApproach":1145,"relations":1147,"seo":1152,"stem":1155,"summary":1156,"visuals":1157,"__hash__":1160},"useCases_en\u002Fuse-cases\u002Fai-oplossing-te-langzaam.md","The AI solution is too slow",{"type":23,"value":1087,"toc":1088},[],{"title":26,"searchDepth":27,"depth":27,"links":1089},[],{"intro":1091,"items":1092},"The solution doesn't have to be instant everywhere. But it does have to feel fast enough for the process in which people actually use it.\n",[1093,1096,1099],{"title":1094,"text":1095},"Faster critical flows","The steps users wait for get attention first.\n",{"title":1097,"text":1098},"Insight into bottlenecks","You know which part of the model, application or integration causes the wait.\n",{"title":1100,"text":1101},"Deliberate performance choices","Speed, quality and cost are weighed against each other instead of being optimised separately.\n",{"intro":1103},"An AI feature that works correctly but needs twenty seconds for every step still feels broken. We find out where the time is really lost and which optimisations make sense.\n",{},"\u002Fuse-cases\u002Fai-oplossing-te-langzaam",{"intro":1107,"causes":1108},"Choosing a faster model helps little if the system meanwhile runs six unnecessary steps.\n",[1109,1112,1115],{"title":1110,"text":1111},"The architecture is fully sequential","Work that could run in parallel or in the background now blocks the user.\n",{"title":1113,"text":1114},"Too much AI for simple tasks","Deterministic logic runs through a model where regular code would be faster.\n",{"title":1116,"text":1117},"Nothing is measured in a targeted way","Without timings and traces you easily end up optimising the wrong part.\n",{"intro":1119,"points":1120},"Speed depends on the whole system. That's why we look at AI, backend, database, integrations and the user flow at the same time.\n",[1121,1124,1127],{"title":1122,"text":1123},"Measure before optimising","Find the bottleneck first, only then make technical choices.\n",{"title":1125,"text":1126},"AI where it adds value","Not every step needs to go through a language model.\n",{"title":1128,"text":1129},"UX counts too","Sometimes the best improvement lies in processing differently or giving smarter feedback.\n",{"intro":1131,"signals":1132},"AI latency often doesn't come from one big problem, but from a whole series of small waits added together.\n",[1133,1136,1139,1142],{"title":1134,"text":1135},"Every step waits for a model","Calls run one after another when that isn't always necessary.\n",{"title":1137,"text":1138},"Agents do much more work than expected","One user action leads to a long chain of tool and model calls.\n",{"title":1140,"text":1141},"Large amounts of context are sent again and again","The same information is fetched and processed over and over.\n",{"title":1143,"text":1144},"External systems slow down the whole flow","One slow API holds back the rest of the user experience.\n",{"intro":1146},"In a Vibe Code Rescue Audit we first measure where time is actually lost. We can then optimise in a targeted way in a Rescue Sprint, without endlessly tinkering with models and prompts.\n",{"buyerTriggerKeys":1148,"primaryOfferKey":73,"offerKeys":1149,"caseKeys":1150,"relatedUseCaseKeys":1151},[35],[73,74],[],[63,56],{"title":1153,"description":1154},"AI solution too slow? | 9to5 software","Slow AI app or agent? 9to5 examines models, tool calls, architecture, caching and integrations and tackles the real bottleneck.\n","use-cases\u002Fai-oplossing-te-langzaam","The functionality works, but users wait so long that the product or process isn't pleasant to use in practice.\n",{"hero":1158,"problem":701,"proof":1159},"sq-bottleneck-pressure","ls-optimize-flow","SBc7Xo_WFVh7eL0GKxJFFPUMszF10AfU3LJbwWTcw7c",{"id":1162,"title":1163,"body":1164,"description":26,"desiredOutcome":1168,"extension":32,"hero":1180,"key":63,"meta":1182,"navigation":38,"path":1183,"problem":1184,"proof":1196,"recognition":1208,"recommendedApproach":1223,"relations":1225,"seo":1230,"stem":1233,"summary":1234,"visuals":1235,"__hash__":1236},"useCases_en\u002Fuse-cases\u002Fai-oplossing-te-duur.md","The AI solution is becoming too expensive",{"type":23,"value":1165,"toc":1166},[],{"title":26,"searchDepth":27,"depth":27,"links":1167},[],{"intro":1169,"items":1170},"You don't simply want to use the cheapest models. You want costs to be in proportion to the value of every user flow.\n",[1171,1174,1177],{"title":1172,"text":1173},"Costs become transparent","You know which actions, users or processes cause the most usage.\n",{"title":1175,"text":1176},"Waste disappears","Unnecessary calls, context and duplicate processing are reduced in a targeted way.\n",{"title":1178,"text":1179},"Growth becomes more predictable","You can better estimate what more usage means financially.\n",{"intro":1181},"A few euros during the prototype is very different from hundreds or thousands of users running the same workflow every day. We find out where the costs come from and how to lower them without optimising the solution into the ground.\n",{},"\u002Fuse-cases\u002Fai-oplossing-te-duur",{"intro":1185,"causes":1186},"AI costs are often not just a pricing problem, but also an architecture and product problem.\n",[1187,1190,1193],{"title":1188,"text":1189},"Expensive models are used everywhere","Simple tasks get the same heavy model choice as complex tasks.\n",{"title":1191,"text":1192},"Workflows do unnecessary work","Calls, retrieval and processing are repeated while results could be reused.\n",{"title":1194,"text":1195},"Costs are not part of observability","So you only see waste when the invoice arrives.\n",{"intro":1197,"points":1198},"The cheapest model call is often the call that turns out not to be needed. That is why we look at the whole system, not just the model providers' price list.\n",[1199,1202,1205],{"title":1200,"text":1201},"Cost and quality","Cheaper doesn't help if usability collapses.\n",{"title":1203,"text":1204},"Architecture over loose tweaks","Routing, caching and workflow design can deliver more than prompt micro-optimisation.\n",{"title":1206,"text":1207},"Measurable results","Optimisation is tied to actual usage per relevant flow.\n",{"intro":1209,"signals":1210},"Cost problems often only become visible when an experiment successfully starts to scale.\n",[1211,1214,1217,1220],{"title":1212,"text":1213},"Model costs grow faster than users","Every new user causes relatively a lot of extra usage.\n",{"title":1215,"text":1216},"Agents make many calls per task","A simple action can quietly involve dozens of model or tool calls.\n",{"title":1218,"text":1219},"Large prompts are sent along every time","Context grows along with the product without it being clear what is really needed.\n",{"title":1221,"text":1222},"Nobody knows the cost per flow","There is a monthly invoice, but hardly any insight into what causes it.\n",{"intro":1224},"With a Vibe Code Rescue Audit we map the costs per relevant flow and the biggest technical causes. We then carry out the most worthwhile optimisations in a Rescue Sprint.\n",{"buyerTriggerKeys":1226,"primaryOfferKey":73,"offerKeys":1227,"caseKeys":1228,"relatedUseCaseKeys":1229},[35],[73,74],[],[62,70],{"title":1231,"description":1232},"AI solution too expensive? Optimise costs | 9to5 software","AI costs (OpenAI, Claude, tokens) rising? 9to5 examines model usage, agent flows, caching and infrastructure to find unnecessary costs.\n","use-cases\u002Fai-oplossing-te-duur","Usage is growing, and with it the model, infrastructure or API costs. The current architecture turns out not to scale well economically.\n",{"hero":1158,"problem":701,"proof":1159},"AvSDQPDw5Ui68JRE-XmUiKA5-WySdoPvNHAXTC42Cu0",{"id":1238,"title":1239,"body":1240,"description":26,"desiredOutcome":1244,"extension":32,"hero":1256,"key":64,"meta":1258,"navigation":38,"path":1259,"problem":1260,"proof":1272,"recognition":1284,"recommendedApproach":1299,"relations":1301,"seo":1306,"stem":1309,"summary":1310,"visuals":1311,"__hash__":1313},"useCases_en\u002Fuse-cases\u002Ftests-evals-monitoring-voor-ai.md","We're missing tests, evals or monitoring",{"type":23,"value":1241,"toc":1242},[],{"title":26,"searchDepth":27,"depth":27,"links":1243},[],{"intro":1245,"items":1246},"You want to change things with confidence because you quickly see what a new prompt, model version or code change actually does.\n",[1247,1250,1253],{"title":1248,"text":1249},"Regressions become visible","Important existing situations are automatically checked again.\n",{"title":1251,"text":1252},"AI quality becomes measurable","You compare changes against relevant examples and criteria.\n",{"title":1254,"text":1255},"Production gives feedback","Errors, latency and unusual behaviour no longer disappear quietly into logs.\n",{"intro":1257},"\"It feels better\" is not a reliable test strategy. If you use AI functionality seriously, you need to be able to see what changes do and when things go wrong in production. We help set up that feedback loop.\n",{},"\u002Fuse-cases\u002Ftests-evals-monitoring-voor-ai",{"intro":1261,"causes":1262},"Traditional software tests remain important, but AI adds behaviour for which there isn't always one exactly correct answer.\n",[1263,1266,1269],{"title":1264,"text":1265},"AI output is probabilistic","Exact string comparisons are not enough for many AI tasks.\n",{"title":1267,"text":1268},"Good examples aren't recorded","Knowledge of what is right and wrong mainly lives in heads and scattered chats.\n",{"title":1270,"text":1271},"Observability stops at technical errors","An HTTP 200 says nothing about the quality of the answer.\n",{"intro":1273,"points":1274},"A measurement layer is only useful to us if it helps engineers see faster whether a change is safe and actually better.\n",[1275,1278,1281],{"title":1276,"text":1277},"Regular tests where possible","Deterministic software simply stays deterministically testable.\n",{"title":1279,"text":1280},"Evals where needed","AI behaviour gets examples and criteria that fit the real use case.\n",{"title":1282,"text":1283},"Production data closes the loop","Monitoring shows which real-world situations need attention.\n",{"intro":1285,"signals":1286},"Without a good feedback loop, an AI product becomes harder and harder to develop responsibly.\n",[1287,1290,1293,1296],{"title":1288,"text":1289},"Testing is mostly manual","After every change someone tries a few known examples.\n",{"title":1291,"text":1292},"Prompt changes are hard to compare","You can't objectively tell whether version B is really better than version A.\n",{"title":1294,"text":1295},"Production errors are discovered by users","The system itself hardly flags deviating behaviour.\n",{"title":1297,"text":1298},"Model updates feel risky","A new provider or model version can change behaviour without clear warning.\n",{"intro":1300},"If the scope is clear, we set up a practical feedback loop in a Rescue Sprint with the right combination of tests, evals, traces and monitoring. If broader research is needed first, we start with an audit.\n",{"buyerTriggerKeys":1302,"primaryOfferKey":74,"offerKeys":1303,"caseKeys":1304,"relatedUseCaseKeys":1305},[35],[74,73],[],[56,58],{"title":1307,"description":1308},"Tests, evals and monitoring for AI | 9to5 software","AI without proper tests, LLM evals or monitoring? 9to5 helps set up measurable quality, regression tests and production observability.\n","use-cases\u002Ftests-evals-monitoring-voor-ai","The AI functionality runs, but nobody can objectively see whether changes make it better, worse or quietly break something.\n",{"hero":329,"problem":701,"proof":1312},"ls-workflow-flow","HLgbDqrGnq-G64ciWMNMDChkwkuEQEyI9a7eCclpOe0",{"id":1315,"title":1316,"body":1317,"description":26,"desiredOutcome":1321,"extension":32,"hero":1333,"key":65,"meta":1335,"navigation":38,"path":1336,"problem":1337,"proof":1349,"recognition":1361,"recommendedApproach":1376,"relations":1378,"seo":1383,"stem":1386,"summary":1387,"visuals":1388,"__hash__":1389},"useCases_en\u002Fuse-cases\u002Fdevelopmentteam-loopt-vast-met-ai.md","Our own development team is technically stuck",{"type":23,"value":1318,"toc":1319},[],{"title":26,"searchDepth":27,"depth":27,"links":1320},[],{"intro":1322,"items":1323},"You don't want to replace the existing team, but to remove the technical blocker so the team can move forward on its own again.\n",[1324,1327,1330],{"title":1325,"text":1326},"The hard part gets solved","Senior engineers tackle exactly the part the team is stuck on.\n",{"title":1328,"text":1329},"Knowledge stays in-house","Your engineers understand the solution and the most important trade-offs.\n",{"title":1331,"text":1332},"No extra dependency","After the rescue, 9to5 doesn't need to remain part of the team permanently.\n",{"intro":1334},"Sometimes you don't need an extra team. You need someone who quickly gets to the bottom of one complicated technical problem, solves it together with your engineers and then leaves again. That is where we like to step in.\n",{},"\u002Fuse-cases\u002Fdevelopmentteam-loopt-vast-met-ai",{"intro":1338,"causes":1339},"AI projects combine several disciplines. Even a strong development team doesn't need to have immediate experience with every specialist part.\n",[1340,1343,1346],{"title":1341,"text":1342},"The problem is multidisciplinary","Software architecture, models, data, integrations and observability all touch.\n",{"title":1344,"text":1345},"Experience with this specific failure mode is missing","Some problems are solved much faster when you've seen them before.\n",{"title":1347,"text":1348},"Internal experimenting gets expensive","After several failed routes, searching further costs more than temporary specialist help.\n",{"intro":1350,"points":1351},"This is how we want to work: bring in senior engineering where it gets hard and leave again once your own team can carry on.\n",[1352,1355,1358],{"title":1353,"text":1354},"Engineers work with engineers","No advisory layer between the problem and the people who actually change the code.\n",{"title":1356,"text":1357},"The existing stack remains the starting point","We don't bring in a new tool or architecture by default.\n",{"title":1359,"text":1360},"Handover from the start","The solution doesn't become needlessly dependent on whoever built it.\n",{"intro":1362,"signals":1363},"A good software team doesn't automatically have all the experience with agents, LLMs, agentic coding (Claude Code, Codex) and production AI in-house.\n",[1364,1367,1370,1373],{"title":1365,"text":1366},"The team keeps returning to the same problem","Different solutions are tried without a structural breakthrough.\n",{"title":1368,"text":1369},"AI-specific choices remain uncertain","Model behaviour, evals, agent architecture or retrieval call for different experience.\n",{"title":1371,"text":1372},"An integration turns out much harder than expected","The problem sits on the boundary between existing software and the new AI functionality.\n",{"title":1374,"text":1375},"A senior hire is disproportionate","You need specific expertise now, but not necessarily permanently.\n",{"intro":1377},"If the problem is sharp enough, we step in right next to the team with a Rescue Sprint. We work in the existing codebase, solve the blocker and transfer the relevant knowledge along the way. If the root cause is unclear, we start with an audit first.\n",{"buyerTriggerKeys":1379,"primaryOfferKey":74,"offerKeys":1380,"caseKeys":1381,"relatedUseCaseKeys":1382},[35],[74,73],[],[66,60],{"title":1384,"description":1385},"Development team stuck with AI | 9to5 software","Development team stuck with AI agents or agentic coding (Claude Code, Codex)? 9to5 steps in temporarily, solves the technical problem and hands it back over.\n","use-cases\u002Fdevelopmentteam-loopt-vast-met-ai","The team builds software just fine, but lacks the specific senior experience to solve a tricky AI problem properly.\n",{"hero":83,"problem":212,"proof":84},"G-CaoUkXnekooXGZnXL-EzJ47Frx3e9u27FGNRjxr3k",{"id":1391,"title":1392,"body":1393,"description":26,"desiredOutcome":1397,"extension":32,"hero":1409,"key":66,"meta":1411,"navigation":38,"path":1412,"problem":1413,"proof":1425,"recognition":1437,"recommendedApproach":1452,"relations":1454,"seo":1459,"stem":1462,"summary":1463,"visuals":1464,"__hash__":1465},"useCases_en\u002Fuse-cases\u002Ftechnische-second-opinion-ai.md","We want an independent technical second opinion",{"type":23,"value":1394,"toc":1395},[],{"title":26,"searchDepth":27,"depth":27,"links":1396},[],{"intro":1398,"items":1399},"You don't want an extra opinion that makes the discussion bigger. You want enough technical clarity to make a better decision yourself.\n",[1400,1403,1406],{"title":1401,"text":1402},"Separate facts from assumptions","You see which technical conclusions are solid and where uncertainty remains.\n",{"title":1404,"text":1405},"Alternatives become visible","You know whether the proposed route really is the only logical option.\n",{"title":1407,"text":1408},"A concrete next decision","After the review you know what to investigate, keep, adjust or drop.\n",{"intro":1410},"Maybe your vendor says everything has to be rebuilt. Maybe your own team thinks it's fine. When a lot of money or risk depends on the next step, an independent technical view can be well worth having.\n",{},"\u002Fuse-cases\u002Ftechnische-second-opinion-ai",{"intro":1414,"causes":1415},"Technical choices are not fully objective. Everyone looks at the same situation from their own experience, interests and preferences.\n",[1416,1419,1422],{"title":1417,"text":1418},"Vendors mainly know their own tools","A proposal can unconsciously steer towards solutions that suit the vendor.\n",{"title":1420,"text":1421},"Existing teams get used to limitations","What feels normal internally can be obvious technical debt to an outsider.\n",{"title":1423,"text":1424},"Big decisions become final too early","Sometimes evidence is still missing for assumptions on which a far-reaching choice is based.\n",{"intro":1426,"points":1427},"A good second opinion doesn't automatically sell a larger follow-up engagement. The outcome may simply be that the existing route is fine.\n",[1428,1431,1434],{"title":1429,"text":1430},"Engineers first","The assessment is made from the perspective of technical feasibility and risk.\n",{"title":1432,"text":1433},"Build, buy or change nothing","We have no stake in one specific technical outcome.\n",{"title":1435,"text":1436},"Follow-up is not required","If the conclusion is clear, the engagement can simply end there.\n",{"intro":1438,"signals":1439},"A second opinion is especially useful when the next technical step is large, expensive or hard to reverse.\n",[1440,1443,1446,1449],{"title":1441,"text":1442},"A vendor wants to rebuild a lot","You want to know whether that's technically necessary or mainly the easiest route for them.\n",{"title":1444,"text":1445},"Internal opinions differ","Different engineers judge the same risks differently.\n",{"title":1447,"text":1448},"The architecture feels more complicated than needed","Many new tools, platforms or layers are being proposed.\n",{"title":1450,"text":1451},"You're facing a big investment","Before continuing, you want to know whether the technical foundation makes sense.\n",{"intro":1453},"The Vibe Code Rescue Audit fits here as an independent technical review. We look at what's actually there, test the most important assumptions and give a concrete technical debrief. Follow-up work is only needed if there is a substantive reason for it.\n",{"buyerTriggerKeys":1455,"primaryOfferKey":73,"offerKeys":1456,"caseKeys":1457,"relatedUseCaseKeys":1458},[35],[73,74],[],[67,70],{"title":1460,"description":1461},"Technical second opinion for an AI project | 9to5 software","Need an independent technical second opinion on an AI solution, architecture or vendor? Senior engineers from 9to5 assess what's there.\n","use-cases\u002Ftechnische-second-opinion-ai","There's already an architecture, solution or proposal, but you want to know whether the technical choices make sense before investing further.\n",{"hero":212,"problem":329,"proof":213},"Mz9vCsA-Sb-E2hgZP7IGG0NRPbBoSbELaFOOyNALsPo",{"id":1467,"title":1468,"body":1469,"description":26,"desiredOutcome":1473,"extension":32,"hero":1485,"key":67,"meta":1487,"navigation":38,"path":1488,"problem":1489,"proof":1501,"recognition":1513,"recommendedApproach":1528,"relations":1530,"seo":1535,"stem":1538,"summary":1539,"visuals":1540,"__hash__":1541},"useCases_en\u002Fuse-cases\u002Fai-project-overnemen-van-andere-leverancier.md","Another vendor left something behind that we can't properly continue with",{"type":23,"value":1470,"toc":1471},[],{"title":26,"searchDepth":27,"depth":27,"links":1472},[],{"intro":1474,"items":1475},"The system should be understandable to more people than just the party that built it.\n",[1476,1479,1482],{"title":1477,"text":1478},"Technical grip regained","Architecture, deployment, data and external dependencies are clear again.\n",{"title":1480,"text":1481},"Risks are prioritised","You know which parts are fine and where work is needed first.\n",{"title":1483,"text":1484},"The project becomes transferable again","Your team or a new vendor can continue development responsibly.\n",{"intro":1486},"The original developer is gone, the documentation is thin and every change starts with the question: does anyone know how this works? We can step in, reconstruct what is there and get you moving again.\n",{},"\u002Fuse-cases\u002Fai-project-overnemen-van-andere-leverancier",{"intro":1490,"causes":1491},"Taking over software is more than getting access to the Git repository.\n",[1492,1495,1498],{"title":1493,"text":1494},"Knowledge sat with people, not in the system","Architecture and operational choices were never documented well enough.\n",{"title":1496,"text":1497},"Ownership is scattered","Code, infrastructure, accounts and data may sit with different parties.\n",{"title":1499,"text":1500},"The technical quality is still unknown","You don't know what merely feels unfamiliar and what is an actual risk.\n",{"intro":1502,"points":1503},"Stepping into existing software is part of senior engineering. You don't need to have built the system yourself to bring structure and grip back into it.\n",[1504,1507,1510],{"title":1505,"text":1506},"We start with what exists","A different vendor is no reason to rebuild right away.\n",{"title":1508,"text":1509},"Code and operations","Deployment, accounts and integrations belong to the handover just as much.\n",{"title":1511,"text":1512},"Avoiding a new dependency","The software should be easier to hand over after our work, not harder.\n",{"intro":1514,"signals":1515},"A handover can be complete on paper while the most important knowledge is still missing.\n",[1516,1519,1522,1525],{"title":1517,"text":1518},"The repository is there, but context is missing","Important choices were never recorded anywhere.\n",{"title":1520,"text":1521},"Deployment is a black box","Nobody knows exactly which steps, accounts or infrastructure are needed.\n",{"title":1523,"text":1524},"External services are scattered","Credentials, providers and admin accounts sit with different people or parties.\n",{"title":1526,"text":1527},"Every change first requires research","Engineers have to reconstruct again and again how functionality fits together.\n",{"intro":1529},"With a Vibe Code Rescue Audit we first reconstruct what is there and what is still missing. Specific technical blockers can then be tackled in a Rescue Sprint.\n",{"buyerTriggerKeys":1531,"primaryOfferKey":73,"offerKeys":1532,"caseKeys":1533,"relatedUseCaseKeys":1534},[35],[73,74],[],[66,69],{"title":1536,"description":1537},"Take over an AI project from another vendor | 9to5 software","Stuck with software from a previous vendor? 9to5 reconstructs architecture and code, maps the risks and makes the project transferable again.\n","use-cases\u002Fai-project-overnemen-van-andere-leverancier","The software exists, but knowledge, documentation or technical grip is missing after another party left.\n",{"hero":83,"problem":470,"proof":84},"rxzyg-E8-PHje_fvva9yRDf64IMLaDvF_b0hlCG-M4Y",{"id":1543,"title":1544,"body":1545,"description":26,"desiredOutcome":1549,"extension":32,"hero":1561,"key":68,"meta":1563,"navigation":38,"path":1564,"problem":1565,"proof":1577,"recognition":1589,"recommendedApproach":1604,"relations":1606,"seo":1611,"stem":1614,"summary":1615,"visuals":1616,"__hash__":1617},"useCases_en\u002Fuse-cases\u002Fai-experiment-heeft-echte-gebruikers.md","Our experiment became unexpectedly successful and suddenly has real users",{"type":23,"value":1546,"toc":1547},[],{"title":26,"searchDepth":27,"depth":27,"links":1548},[],{"intro":1550,"items":1551},"You don't want to slow down growth, but you also don't want to wait until the first serious outage, security flaw or cost explosion forces you to act.\n",[1552,1555,1558],{"title":1553,"text":1554},"The biggest risks under control","Security, data, deployment and reliability get the attention that real use requires.\n",{"title":1556,"text":1557},"Growth stays manageable","More users does not automatically mean proportionally more technical problems.\n",{"title":1559,"text":1560},"The experiment can become a product","The most important prototype shortcuts are deliberately kept or replaced.\n",{"intro":1562},"Good news: people want to use your product. Less good news: the software was never built for that. We help make sure unexpected success doesn't turn into a technical incident.\n",{},"\u002Fuse-cases\u002Fai-experiment-heeft-echte-gebruikers",{"intro":1566,"causes":1567},"Prototype choices were made for speed of learning. Success suddenly changes the requirements the same software has to meet.\n",[1568,1571,1574],{"title":1569,"text":1570},"Operational requirements came later","Monitoring, backups, support and scalability were, logically enough, no priority during the experiment.\n",{"title":1572,"text":1573},"The user population changes","Real users do things the original test team never thought of.\n",{"title":1575,"text":1576},"Business value raises technical risk","As soon as people depend on the system, downtime has a different impact.\n",{"intro":1578,"points":1579},"We don't make everything enterprise-grade by default. The point is to work out which engineering you need now to keep growing responsibly.\n",[1580,1583,1586],{"title":1581,"text":1582},"Risk-driven professionalisation","We first tackle what really threatens users, data or continuity.\n",{"title":1584,"text":1585},"Keep prototype knowledge","Working parts are not replaced just because they started as an experiment.\n",{"title":1587,"text":1588},"Ready for the next phase","The technical foundation grows along with actual product use.\n",{"intro":1590,"signals":1591},"The technology is often fine, but built for a very different scale and expectation.\n",[1592,1595,1598,1601],{"title":1593,"text":1594},"Support questions suddenly come in","Users expect availability and predictable behaviour.\n",{"title":1596,"text":1597},"Accounts and permissions become important","What was first one internal experiment now has different users and roles.\n",{"title":1599,"text":1600},"Costs start to stand out","Model and infrastructure usage suddenly grows along with real use.\n",{"title":1602,"text":1603},"You're afraid to make big changes","The software is now important enough that downtime or data loss hurts.\n",{"intro":1605},"A Vibe Code Rescue Audit first helps determine what is truly urgent now. We can then choose between focused Rescue Sprints and a larger Build & Integrate engagement to professionalise the product further.\n",{"buyerTriggerKeys":1607,"primaryOfferKey":73,"offerKeys":1608,"caseKeys":1609,"relatedUseCaseKeys":1610},[35],[73,75],[],[55,69],{"title":1612,"description":1613},"AI experiment suddenly has real users | 9to5 software","AI experiment unexpectedly successful? 9to5 helps you scale your prototype safely, reliably and maintainably to real use.\n","use-cases\u002Fai-experiment-heeft-echte-gebruikers","What started as an experiment is actually being used. Now the technology has to mature faster than originally planned.\n",{"hero":471,"problem":1158,"proof":624},"xBug-F1_Gp1DHk40W3mSx2FhRCmjWWw1lcKDGuJ_ghA",{"id":1619,"title":1620,"body":1621,"description":26,"desiredOutcome":1625,"extension":32,"hero":1637,"key":69,"meta":1639,"navigation":38,"path":1640,"problem":1641,"proof":1653,"recognition":1665,"recommendedApproach":1680,"relations":1682,"seo":1687,"stem":1690,"summary":1691,"visuals":1692,"__hash__":1693},"useCases_en\u002Fuse-cases\u002Fprototype-naar-onderhoudbaar-softwareproduct.md","We want to go from prototype to a maintainable software product",{"type":23,"value":1622,"toc":1623},[],{"title":26,"searchDepth":27,"depth":27,"links":1624},[],{"intro":1626,"items":1627},"Starting over isn't a given. What you want is a technical foundation on which a normal development team can keep building responsibly.\n",[1628,1631,1634],{"title":1629,"text":1630},"Understandable architecture","It's clear where responsibilities and business logic belong.\n",{"title":1632,"text":1633},"A safe development cycle","Tests, deployment and error handling support frequent change.\n",{"title":1635,"text":1636},"Less knowledge in people's heads","Important technical choices and processes are sufficiently documented.\n",{"intro":1638},"A prototype mainly has to answer one question quickly: does the idea work? A software product also has to stay understandable, survive changes and be maintainable by more than one person.\n",{},"\u002Fuse-cases\u002Fprototype-naar-onderhoudbaar-softwareproduct",{"intro":1642,"causes":1643},"Prototype code is optimised for speed of learning. Product software also has to be optimised for change.\n",[1644,1647,1650],{"title":1645,"text":1646},"Architecture lagged behind product development","The prototype grew organically while requirements kept changing.\n",{"title":1648,"text":1649},"Temporary assumptions became structural","Choices that were logical for an experiment now determine how the whole product works.\n",{"title":1651,"text":1652},"Maintainability was not yet a success criterion","Tests, documentation and clear boundaries delivered little direct value during the first validation.\n",{"intro":1654,"points":1655},"9to5 software combines AI with regular senior software engineering. That is exactly what matters when a smart prototype has to become a durable product.\n",[1656,1659,1662],{"title":1657,"text":1658},"Keep what works","A rewrite is one possible choice, not the default answer.\n",{"title":1660,"text":1661},"Maintainability is practical","We focus on changing, testing and releasing more safely.\n",{"title":1663,"text":1664},"Built for the next team","Software must stay understandable without its original builder too.\n",{"intro":1666,"signals":1667},"A lot of prototype code isn't bad at all. You just see more and more clearly where speed was chosen over maintainability.\n",[1668,1671,1674,1677],{"title":1669,"text":1670},"Temporary solutions have become permanent","Scripts, hardcoded values and shortcuts are now used by real flows.\n",{"title":1672,"text":1673},"New features take more and more effort","The existing structure no longer fits well with what the product has become.\n",{"title":1675,"text":1676},"Only one person understands critical parts","Knowledge transfer becomes hard and changes feel risky.\n",{"title":1678,"text":1679},"Releases are largely manual work","Deployment and configuration need steps that are easily forgotten or done wrong.\n",{"intro":1681},"With a Vibe Code Rescue Audit we determine what forms a good foundation and what really holds development back. Build & Integrate can then carry out the larger professionalisation without automatically rebuilding the whole product.\n",{"buyerTriggerKeys":1683,"primaryOfferKey":73,"offerKeys":1684,"caseKeys":1685,"relatedUseCaseKeys":1686},[35],[73,75],[],[55,54],{"title":1688,"description":1689},"From prototype to maintainable software product | 9to5 software","Developing a prototype into maintainable software? 9to5 helps get architecture, code, tests, deployment and documentation ready for long-term use.\n","use-cases\u002Fprototype-naar-onderhoudbaar-softwareproduct","The core works, but the software has to be built so that a team can still change it safely in six or twelve months.\n",{"hero":471,"problem":83,"proof":624},"DfBDuBl23pxVzfnTmE43TF-OjYwrcBmYDpM95_zySzw",{"id":1695,"title":1696,"body":1697,"description":26,"desiredOutcome":1701,"extension":32,"hero":1713,"key":70,"meta":1715,"navigation":38,"path":1716,"problem":1717,"proof":1729,"recognition":1740,"recommendedApproach":1755,"relations":1757,"seo":1762,"stem":1765,"summary":1766,"visuals":1767,"__hash__":1768},"useCases_en\u002Fuse-cases\u002Fweg-bij-ai-platform-of-leverancier.md","We want to leave an AI platform or vendor but are technically stuck",{"type":23,"value":1698,"toc":1699},[],{"title":26,"searchDepth":27,"depth":27,"links":1700},[],{"intro":1702,"items":1703},"You want the vendor to become a technical choice again, instead of a condition for your product's survival.\n",[1704,1707,1710],{"title":1705,"text":1706},"Insight into real dependencies","You know which data, functionality and workflows are tied specifically to the vendor.\n",{"title":1708,"text":1709},"A realistic migration route","Not everything has to be replaced in one risky big bang.\n",{"title":1711,"text":1712},"More technical freedom of choice","The new architecture makes future changes easier.\n",{"intro":1714},"A platform that gave you speed at the start can later hold you back. Maybe the costs are too high, functionality is missing or you want more control. We find out how deep the dependency really goes and how to get out in a controlled way.\n",{},"\u002Fuse-cases\u002Fweg-bij-ai-platform-of-leverancier",{"intro":1718,"causes":1719},"Vendor lock-in rarely sits in a single API call. Dependencies can be spread across data, code, infrastructure and operational processes.\n",[1720,1723,1726],{"title":1721,"text":1722},"Convenience hides technical choices","Managed platforms combine several functions, so the boundaries only become visible when you want to leave.\n",{"title":1724,"text":1725},"Proprietary functionality grows along","The longer the platform is used, the more specific features become part of the product.\n",{"title":1727,"text":1728},"Migration risk becomes too big","Eventually nobody dares to change anything because it's unclear what will be affected.\n",{"intro":1730,"points":1731},"We're not tied to one model provider or AI platform. The technology should serve the product, not the other way around.\n",[1732,1735,1737],{"title":1733,"text":1734},"Independent technical advice","We assess keeping, migrating and replacing as real alternatives.\n",{"title":413,"text":1736},"We can build adapters and migration layers where a phased switch is smarter.\n",{"title":1738,"text":1739},"Deliberately reduce lock-in","The new situation should offer more freedom of choice than the old one.\n",{"intro":1741,"signals":1742},"Lock-in often only becomes visible the moment you actually want to change something.\n",[1743,1746,1749,1752],{"title":1744,"text":1745},"Exporting data turns out to be hard","Important information sits in proprietary storage or formats.\n",{"title":1747,"text":1748},"Workflows use platform-specific features","The functionality isn't easily available one-to-one elsewhere.\n",{"title":1750,"text":1751},"Costs or terms are changing","The platform is no longer as attractive as when you started.\n",{"title":1753,"text":1754},"Nobody knows how big a migration will be","Your own code and vendor functionality have become intertwined over time.\n",{"intro":1756},"With a Vibe Code Rescue Audit we first map dependencies and migration risks. Build & Integrate can then carry out a controlled migration if that turns out to be the wisest route.\n",{"buyerTriggerKeys":1758,"primaryOfferKey":73,"offerKeys":1759,"caseKeys":1760,"relatedUseCaseKeys":1761},[35],[73,75],[],[63,66],{"title":1763,"description":1764},"Leave an AI platform or vendor | 9to5 software","Facing vendor lock-in with OpenAI or another AI platform? 9to5 maps dependencies and migrates models, data and workflows in a controlled way.\n","use-cases\u002Fweg-bij-ai-platform-of-leverancier","The current vendor, tool or AI platform no longer fits, but data, workflows or technical dependencies make switching hard.\n",{"hero":470,"problem":212,"proof":348},"UI0SAkxx7iW6IyBmiPQtsI2VrQR5eXizTsvaZCFyAP4",{"id":1770,"title":1771,"body":1772,"description":26,"desiredOutcome":1776,"extension":32,"hero":1788,"key":71,"meta":1790,"navigation":38,"path":1791,"problem":1792,"proof":1804,"recognition":1816,"recommendedApproach":1831,"relations":1833,"seo":1838,"stem":1841,"summary":1842,"visuals":1843,"__hash__":1844},"useCases_en\u002Fuse-cases\u002Fvibe-code-project-vastgelopen.md","We've vibe-coded ourselves into a corner",{"type":23,"value":1773,"toc":1774},[],{"title":26,"searchDepth":27,"depth":27,"links":1775},[],{"intro":1777,"items":1778},"You don't want to stop AI-assisted development. You want to be able to build again without every change introducing a new technical risk.\n",[1779,1782,1785],{"title":1780,"text":1781},"Grip on the codebase again","You understand how the most important parts fit together and where you can safely make changes.\n",{"title":1783,"text":1784},"Changing without roulette","New features no longer unexpectedly break other parts of the application.\n",{"title":1786,"text":1787},"Able to carry on yourself","The technical foundation and way of working are clear enough to build on independently with AI tools again.\n",{"intro":1789},"At first it felt like you had superpowers, with a new feature every evening. Then came the point where the AI mainly found new ways to move old problems around. Often you don't have to redo everything.\n",{},"\u002Fuse-cases\u002Fvibe-code-project-vastgelopen",{"intro":1793,"causes":1794},"AI coding tools are very good at momentum. But without clear architecture, tests and technical boundaries, that momentum can eventually work against you.\n",[1795,1798,1801],{"title":1796,"text":1797},"Local solutions pile up","Every prompt solves the problem of that moment, but doesn't automatically take the architecture of the whole into account.\n",{"title":1799,"text":1800},"Technical feedback is missing","Without good tests, types, conventions and deployment checks, the AI only notices late that a change causes damage somewhere else.\n",{"title":1802,"text":1803},"The codebase itself becomes the bottleneck","More and more context is needed to make a change properly. So fast building slowly becomes slow and unpredictable again.\n",{"intro":1805,"points":1806},"This is where senior software engineering makes the difference: understanding an existing codebase, reducing technical risks and making sure your own team can continue independently afterwards.\n",[1807,1810,1813],{"title":1808,"text":1809},"We're not against vibe coding","Building yourself with AI can actually be very effective when the technical foundation stays good enough.\n",{"title":1811,"text":1812},"We don't automatically start over","First determine what is usable, only then decide what really has to change.\n",{"title":1814,"text":1815},"The idea is that you build again","A rescue has succeeded when you don't need 9to5 for every next feature afterwards.\n",{"intro":1817,"signals":1818},"Vibe coding rarely gets stuck all at once. First small technical problems pile up, until more and more prompts are needed to change anything.\n",[1819,1822,1825,1828],{"title":1820,"text":1821},"One fix breaks something else","A seemingly small change suddenly causes errors in places you didn't touch.\n",{"title":1823,"text":1824},"The AI keeps suggesting bigger refactors","According to the coding tool, even simple changes seem to need a major rework of the code.\n",{"title":1826,"text":1827},"Nobody knows what leads anymore","Several implementations, old files and half-replaced solutions exist side by side.\n",{"title":1829,"text":1830},"You hardly dare touch working code","The project still works, but every change feels like you might knock over something important.\n",{"intro":1832},"Don't keep prompting blindly, and don't automatically rebuild everything either. In a Vibe Code Rescue Audit we separate the good parts from temporary solutions and structural risks. You then get a concrete route: what can stay, what needs improving and what really doesn't need redoing.\n",{"buyerTriggerKeys":1834,"primaryOfferKey":73,"offerKeys":1835,"caseKeys":1836,"relatedUseCaseKeys":1837},[35],[73,74],[],[54,57],{"title":1839,"description":1840},"Stuck with vibe coding? | 9to5 software","Vibe-coded yourself into a corner with Lovable, Cursor, Bolt or Replit? 9to5 analyses your codebase, finds the blockers and helps you build on responsibly.\n","use-cases\u002Fvibe-code-project-vastgelopen","With AI coding tools it went incredibly fast at first. Now new changes mainly make the project more complicated and you no longer know how to get out.\n",{"hero":83,"problem":329,"proof":84},"ao7p7vb_k93FD1bj9inbhcKsmZaEIvN9j4_hOr_0TkA",[1846,1909],{"id":1847,"title":1848,"approach":1849,"body":1864,"cta":1868,"description":26,"extension":32,"hero":1871,"key":1873,"label":1874,"meta":1875,"navigation":38,"path":1876,"proof":1877,"relations":1889,"seo":1902,"stem":1905,"summary":1906,"visuals":1907,"__hash__":1908},"buyerTriggers_en\u002Fbuyer-triggers\u002Fstarten-met-ai.md","We want to get started with AI, but aren't sure how.",{"title":1850,"steps":1851},"From a challenge to a sound technical approach.",[1852,1855,1858,1861],{"title":1853,"text":1854},"Define the problem","We establish which decision you need to make, which process you want to improve and what a good result would look like.\n",{"title":1856,"text":1857},"Understand the context and systems","We look at existing software, data, integrations, users and technical constraints.\n",{"title":1859,"text":1860},"Assess the options","We compare AI, automation, existing software, integrations and custom development in terms of feasibility, risk and value.\n",{"title":1862,"text":1863},"Deliver a decision and a way forward","You get a concrete technical direction, the main risks and a logical next step.\n",{"type":23,"value":1865,"toc":1866},[],{"title":26,"searchDepth":27,"depth":27,"links":1867},[],{"title":1869,"body":1870},"Not sure whether AI is the right solution here?","Briefly describe the idea or process you have in mind. We'll quickly tell you what looks technically promising and what a sensible first step would be.\n",{"intro":1872},"You might have an idea, a process that could work better or simply a sense that AI could do more for you. We help you work out what looks promising, where the risks lie and whether AI is actually the right solution.\n","starten-met-ai","Start with AI",{},"\u002Fbuyer-triggers\u002Fstarten-met-ai",{"intro":1878,"points":1879},"We don't advise from a fixed preference for AI or custom software. Because we build software ourselves, we can honestly assess what is technically feasible and worthwhile, and what you are better off not building.\n",[1880,1883,1886],{"title":1881,"text":1882},"Engineers look at the real system","We look at the AI idea, but also at the data, existing software, integrations and users.\n",{"title":1884,"text":1885},"AI is a tool, not the default answer","Automation, existing software, an integration or building nothing are all serious options too.\n",{"title":1887,"text":1888},"Advice and implementation from the same people","If building turns out to make sense, we can carry out the next steps ourselves. That keeps our advice realistic.\n",{"useCaseKeys":1890,"offerKeys":1899,"caseKeys":1901},[1891,1892,1893,1894,1895,1896,1897,1898],"processen-automatiseren-met-ai","interne-ai-assistent-met-eigen-kennis","ai-idee-technisch-haalbaar","is-ai-nodig","ai-betrouwbaar-genoeg","zelf-bouwen-of-software-kopen","ai-toevoegen-aan-bestaand-product","vibe-coding-goed-beginnen",[428,1900,376],"vibe-code-kickstart",[],{"title":1903,"description":1904},"Getting started with AI? From idea to technical approach | 9to5 software","Help with a serious AI challenge. An AI agent, ChatGPT, Claude or conventional software? We assess what is feasible, what not to build and the next step.\n","buyer-triggers\u002Fstarten-met-ai","We see opportunities for AI, but first want to know what makes sense, what is feasible and what is technically sound.\n",{"hero":212,"proof":213,"approach":213},"oMS04pD9lTBstiI46yn57yKInU-_urtigoQYGlj0PWA",{"id":1910,"title":1911,"approach":1912,"body":1927,"cta":1931,"description":26,"extension":32,"hero":1934,"key":421,"label":1936,"meta":1937,"navigation":38,"path":1938,"proof":1939,"relations":1951,"seo":1957,"stem":1960,"summary":1961,"visuals":1962,"__hash__":1964},"buyerTriggers_en\u002Fbuyer-triggers\u002Fai-laten-bouwen.md","We know what we want. Now it needs to be built.",{"title":1913,"steps":1914},"From a clear scope to something that really works.",[1915,1918,1921,1924],{"title":1916,"text":1917},"Define the scope and success criteria","We make it clear what needs to work, who it is for and when it is good enough to go live.\n",{"title":1919,"text":1920},"Set up the architecture and integrations","We determine how AI, data, existing software, permissions and integrations should work together technically.\n",{"title":1922,"text":1923},"Build and test","We implement the difficult part and test more than the happy path, including reliability, errors and human oversight where needed.\n",{"title":1925,"text":1926},"Bring it into production and hand over","We launch it properly, document the important decisions and make sure your team can carry on independently.\n",{"type":23,"value":1928,"toc":1929},[],{"title":26,"searchDepth":27,"depth":27,"links":1930},[],{"title":1932,"body":1933},"Know what needs to be built?","Explain what you need and which systems it should work with. We'll quickly tell you how to build it properly and what a logical first step would be.\n",{"intro":1935},"You know what needs to be built. We help you build it properly, connect it to your existing systems and get it running reliably in production, using Claude, ChatGPT\u002FOpenAI, Codex or MCP where appropriate. We step in, build, document and hand over, without an open-ended staffing commitment.\n","Build with AI",{},"\u002Fbuyer-triggers\u002Fai-laten-bouwen",{"intro":1940,"points":1941},"AI is only really useful when it works with existing software, data and business processes. We build that as software engineering, not as a standalone demo.\n",[1942,1945,1948],{"title":1943,"text":1944},"Integrations are central to our work","We treat APIs, databases, permissions and existing systems as core parts of the solution, not extras around the AI.\n",{"title":1946,"text":1947},"Reliable beyond the happy path","We also test errors, edge cases and human oversight where needed.\n",{"title":1949,"text":1950},"Production and handover are part of the job","The work is done when the solution is live, documented and manageable for your team.\n",{"useCaseKeys":1952,"offerKeys":1955,"caseKeys":1956},[357,358,359,360,423,424,1953,1954,425,426],"ai-prototype-laten-bouwen","product-goed-opzetten-voor-ai-development",[75,376,1900],[],{"title":1958,"description":1959},"AI agent development and integration | 9to5 software","Have senior engineers build an AI agent, MCP integration, assistant or automation with Claude, ChatGPT\u002FOpenAI or Codex. From scope to production.\n","buyer-triggers\u002Fai-laten-bouwen","The problem and direction are clear. Now it needs to be built properly, integrated and brought into production.\n",{"hero":471,"proof":1963,"approach":1963},"ls-system-assemble","ShT4oNEyuY5KbwbXvOuC2vEuCmku084FYvyfv89UAyM",[],{"id":1967,"title":1968,"extension":1969,"items":1970,"meta":1983,"stem":1984,"__hash__":1985},"why9to5_en\u002Fen\u002Fshared\u002Fwhy-9to5.yml","Why 9to5?","yml",[1971,1974,1977,1980],{"label":1972,"value":1973},"Senior engineering","AI + software engineering",{"label":1975,"value":1976},"Existing software","We step in where it gets complicated",{"label":1978,"value":1979},"Production-ready","Built, tested and reliably deployed",{"label":1981,"value":1982},"Handover","Documented, so you can carry on independently",{},"en\u002Fshared\u002Fwhy-9to5","_iz4GBKRty2Gu5JZKf6tSCRXdwX0sju7lRprYEzQvNE",1792049913987]