[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"sanity-V0BhhREjmas3OElTwAzo8aeyK3lxSDqu9vEzYLFWpFA":3},{"data":4,"sourceMap":-1},{"meta":5,"resource":9},{"sidebarCta":6},{"body":7,"title":8},"Subscribe to receive Stack Internal content around knowledge sharing, collaboration, and AI.","Stay updated",{"_createdAt":10,"_id":11,"_rev":12,"_system":13,"_type":16,"_updatedAt":17,"body":18,"category":154,"chapters":157,"displayContents":981,"image":982,"product":986,"publishedAt":1042,"resourceType":1043,"slug":1047,"subcategory":1049,"title":1051,"visible":981},"2026-10-26T15:57:16Z","47d538ea-60fa-4561-aaac-08f7c6194069","t1dTUgrS4HcS7e8NDTkJRg",{"base":14},{"id":11,"rev":15},"G7USs0ErMAs1wCqV757MbB","resource","2026-11-13T21:07:52Z",[19,42,50,58,65,73,83,91,99,106,115,123,131,139,146],{"_key":20,"_type":21,"children":22,"markDefs":37,"style":41},"824a86d90d2a","block",[23,28,33],{"_key":24,"_type":25,"marks":26,"text":27},"3690377c5354","span",[],"Developer confidence in AI has increased over the last year, according to our ",{"_key":29,"_type":25,"marks":30,"text":32},"260194e0b92d",[31],"6bbba3dbccfb","2026 Developer Survey.",{"_key":34,"_type":25,"marks":35,"text":36},"75d6ae2712b1",[]," This is great news from an adoption standpoint. However, this also comes with some caveats and opportunities for improvement.",[38],{"_key":31,"_type":39,"href":40},"link","https:\u002F\u002Fsurvey.stackoverflow.co\u002F2026","normal",{"_key":43,"_type":21,"children":44,"markDefs":49,"style":41},"d0c09da6cc8b",[45],{"_key":46,"_type":25,"marks":47,"text":48},"e1e3614f595a",[],"",[],{"_key":51,"_type":21,"children":52,"markDefs":57,"style":41},"e01ee1f1b972",[53],{"_key":54,"_type":25,"marks":55,"text":56},"94b7e66b5dd3",[],"Practitioners tend to lean away from using AI when it comes to tasks that require more subjective judgment and human communication. This is entirely understandable. If a developer uses incorrect information, it can have big consequences. They want to do their job well, and that means being careful with the data they use. The possibility of incorrect AI outputs make it hard for teams to move from individual use to scaling agentic workflows across teams.",[],{"_key":59,"_type":21,"children":60,"markDefs":64,"style":41},"d98ed5130990",[61],{"_key":62,"_type":25,"marks":63,"text":48},"1dabe9594a15",[],[],{"_key":66,"_type":21,"children":67,"markDefs":72,"style":41},"a867d4e6321b",[68],{"_key":69,"_type":25,"marks":70,"text":71},"34a445ad337a",[],"The 2026 Developer Survey released in October surveyed over 30K technologists across 169 countries with 103 questions in five categories: work, AI, technology, community, and knowledge. In this guide, we will uncover some of the key data from this survey around:",[],{"_key":74,"_type":21,"children":75,"level":80,"listItem":81,"markDefs":82,"style":41},"0fbc4a600baa",[76],{"_key":77,"_type":25,"marks":78,"text":79},"5301b389fe78",[],"What unreliable AI really costs",1,"bullet",[],{"_key":84,"_type":21,"children":85,"level":80,"listItem":81,"markDefs":90,"style":41},"0d44dff7d9d2",[86],{"_key":87,"_type":25,"marks":88,"text":89},"c32b1ac46bed",[],"Why the fix is providing better context",[],{"_key":92,"_type":21,"children":93,"level":80,"listItem":81,"markDefs":98,"style":41},"00105c28c9ae",[94],{"_key":95,"_type":25,"marks":96,"text":97},"dbd10d6a2618",[],"Why the true AI return on investment is outcomes",[],{"_key":100,"_type":21,"children":101,"markDefs":105,"style":41},"ecdc2c8c55c8",[102],{"_key":103,"_type":25,"marks":104,"text":48},"0e9c9522df2f",[],[],{"_key":107,"_type":21,"children":108,"markDefs":114,"style":41},"4a631292f8ac",[109],{"_key":110,"_type":25,"marks":111,"text":113},"98ff1f268d6b",[112],"strong","Key takeaways:",[],{"_key":116,"_type":21,"children":117,"level":80,"listItem":81,"markDefs":122,"style":41},"a4c6e3632959",[118],{"_key":119,"_type":25,"marks":120,"text":121},"933371d07181",[],"Unreliable AI outputs have a hidden tax on businesses.",[],{"_key":124,"_type":21,"children":125,"level":80,"listItem":81,"markDefs":130,"style":41},"e7daf565c3cc",[126],{"_key":127,"_type":25,"marks":128,"text":129},"46371da7550b",[],"A lack of foundational context is a core problem for teams using AI. Practitioners have more confidence in answers from verifiable sources.",[],{"_key":132,"_type":21,"children":133,"level":80,"listItem":81,"markDefs":138,"style":41},"ec84ea4c3986",[134],{"_key":135,"_type":25,"marks":136,"text":137},"f9f563213505",[],"Outcomes are the true value of AI integration that leaders should focus on, and better outcomes start with shared organizational knowledge that can be verified.",[],{"_key":140,"_type":21,"children":141,"markDefs":145,"style":41},"f1ad6b949b0b",[142],{"_key":143,"_type":25,"marks":144,"text":48},"0f860a6a19da",[],[],{"_key":147,"_type":21,"children":148,"markDefs":153,"style":41},"792056e43226",[149],{"_key":150,"_type":25,"marks":151,"text":152},"e64543ffce12",[],"Each section concludes with practical guidance for every key buyer or decision maker from CIOs and Engineering VPs to CISOs and FinOps leaders. The goal is to help improve the practitioner experience with AI integration, allowing teams to use the tools they prefer, connect with the subject matter experts, and set goals focused on business outcomes.",[],{"_ref":155,"_type":156},"ef0e02bc-df9c-4b7c-b754-59031956a409","reference",[158,202,300,446,681,870],{"_key":159,"_type":160,"body":161,"sidebarCta":197,"slug":198,"title":201},"638b48392253","chapter",[162,170,177,183,190],{"_key":163,"_type":21,"children":164,"markDefs":169,"style":41},"e68852600399",[165],{"_key":166,"_type":25,"marks":167,"text":168},"e1e0a02ac20d",[],"The 2026 Developer Survey shows that the tech workforce is using AI: 77% use AI tools daily, and 55% of daily AI users used AI at least two or more hours per day in their work. But despite leadership hoping it will increase team productivity, teams are slowed down by incorrect or inconsistent outputs. As a result, developers end up using valuable time and resources filling in the missing organizational context, like who owns an answer or whether that answer is current and has been validated, to get the information they need to keep working. Someone verifying every output due to missing context is where the anticipated value of AI can leak in an enterprise environment. It’s ultimately a verification and context problem.",[],{"_key":171,"_type":21,"children":172,"markDefs":176,"style":41},"b8ad3c29efce",[173],{"_key":174,"_type":25,"marks":175,"text":48},"7ae34822c26e",[112],[],{"_key":178,"_type":179,"alt":180,"asset":181},"60055a19bc20","image","Frequency of AI tool usage across four categories: AI coding assistants, general-purpose chat tools, AI agents or automated workflows, and internal tools.",{"_ref":182,"_type":156},"image-f3bc86b5b53766937cc1f99841cbdbec1f85e1ca-883x578-png",{"_key":184,"_type":21,"children":185,"markDefs":189,"style":41},"0aa2d104fa0c",[186],{"_key":187,"_type":25,"marks":188,"text":48},"2e5f2a0041c3",[],[],{"_key":191,"_type":21,"children":192,"markDefs":196,"style":41},"621736d13e3e",[193],{"_key":194,"_type":25,"marks":195,"text":48},"71b6c9356b49",[],[],[],{"_type":199,"current":200},"slug","confidence-in-ai-outputs","Confidence in AI outputs",{"_key":203,"_type":160,"body":204,"sidebarCta":296,"slug":297,"title":299},"8c085654213a",[205,224,231,239,246,254,261,266,273,281,288],{"_key":206,"_type":21,"children":207,"markDefs":221,"style":41},"c9da94afb446",[208,212,217],{"_key":209,"_type":25,"marks":210,"text":211},"fa9820ff1aac",[],"The cost of AI hallucinations or near-miss outputs doesn’t show up as a clear line item. ",{"_key":213,"_type":25,"marks":214,"text":216},"ab4fad21ecac",[215],"3456078b0561","Value is lost through hours",{"_key":218,"_type":25,"marks":219,"text":220},"ad24248b55e9",[]," spent rechecking, reworking, and re-prompting on outputs.",[222],{"_key":215,"_type":39,"href":223},"https:\u002F\u002Fsurvey.stackoverflow.co\u002F2026\u002Fknowledge\u002Fdata\u002Fknowledge-a",{"_key":225,"_type":21,"children":226,"markDefs":230,"style":41},"ecbf6a69d36e",[227],{"_key":228,"_type":25,"marks":229,"text":48},"36e5ae2b76c1",[],[],{"_key":232,"_type":21,"children":233,"markDefs":238,"style":41},"a822428b8211",[234],{"_key":235,"_type":25,"marks":236,"text":237},"ff23fd50d86c",[],"The data this year shows that a majority indicate trusting AI at various levels, and 48% specifically trust it when they can easily verify the AI output. When rating the importance of being able to quickly check these sources while working, the top-rated sources were code repos, customer or sales support docs, and internal documentation.",[],{"_key":240,"_type":21,"children":241,"markDefs":245,"style":41},"c4f95ab75c71",[242],{"_key":243,"_type":25,"marks":244,"text":48},"2ae75c5e46b9",[],[],{"_key":247,"_type":21,"children":248,"markDefs":253,"style":41},"218e327f7c08",[249],{"_key":250,"_type":25,"marks":251,"text":252},"1d370f3755a0",[],"This makes sense. AI tools can increase efficiency by handling routine work well enough to make it worth keeping in a workflow. But the outputs aren't quite reliable enough for harder tasks when there is a risk of damaging or expensive errors. This is why usage continues to increase while confidence remains conditional.",[],{"_key":255,"_type":21,"children":256,"markDefs":260,"style":41},"f4d17c244260",[257],{"_key":258,"_type":25,"marks":259,"text":48},"83fdaa281116",[],[],{"_key":262,"_type":179,"alt":263,"asset":264},"5935055a87a1","Top AI mistakes: Plausible but wrong code, hard-to-review generated code, missing context, hallucinated package\u002Fconfig, and incorrect\u002Fstale knowledge.",{"_ref":265,"_type":156},"image-2afaf4bb77dbaee6de672f59f3780df0329a0045-883x534-png",{"_key":267,"_type":21,"children":268,"markDefs":272,"style":41},"5060900962f7",[269],{"_key":270,"_type":25,"marks":271,"text":48},"f56b6c58c647",[],[],{"_key":274,"_type":21,"children":275,"markDefs":280,"style":41},"adf34ffd8110",[276],{"_key":277,"_type":25,"marks":278,"text":279},"b62d250a0924",[],"The top disruptive issues caused by AI in the workplace are plausible-but-wrong code, hard-to-review code, and missing context. Poor debugging paths and security or privacy issues stemming from AI are additional sources of lost team time due to unreliable AI responses. Further complicating things for teams, incorrect outputs that are prematurely relied on often get caught later in the review process or after something has already been shipped and breaks, causing more rework and delays.",[],{"_key":282,"_type":21,"children":283,"markDefs":287,"style":41},"9f804046ede9",[284],{"_key":285,"_type":25,"marks":286,"text":48},"0ef6abf4deeb",[],[],{"_key":289,"_type":21,"children":290,"markDefs":295,"style":41},"00d8b5c6ba12",[291],{"_key":292,"_type":25,"marks":293,"text":294},"88857735c27e",[],"These AI frustrations eventually become a budget problem. When incorrect outputs, hallucinations, and slow reviews pile up, increased work time takes away from the productivity gains an AI tool was meant to deliver. When net gains can’t be shown, it’s hard to justify value or keep working on a project that isn’t leading to the desired outcome fast enough.",[],[],{"_type":199,"current":298},"what-unreliable-ai-actually-costs","What unreliable AI actually costs",{"_key":301,"_type":160,"body":302,"sidebarCta":442,"slug":443,"title":445},"65777158b3b6",[303,311,318,326,331,339,346,354,361,366,373,380,388,395,403,410,418,430],{"_key":304,"_type":21,"children":305,"markDefs":310,"style":41},"6c03f6b05fa5",[306],{"_key":307,"_type":25,"marks":308,"text":309},"57bdbc3a9b16",[],"If you ask a developer what slows them down the most, it isn't writing code. It’s the time it takes to track down the context they need to find the correct information when they notice an inaccurate AI response. This is the context tax that isn’t budgeted for. AI is supposed to save employees time, but for a lot of enterprise teams, the time they are losing because of unreliable AI results is spent tracking down verifiable information.",[],{"_key":312,"_type":21,"children":313,"markDefs":317,"style":41},"3b81781afe9d",[314],{"_key":315,"_type":25,"marks":316,"text":48},"80b910fdaf98",[],[],{"_key":319,"_type":21,"children":320,"markDefs":325,"style":41},"fcd001e78cc0",[321],{"_key":322,"_type":25,"marks":323,"text":324},"f9557fceef2b",[],"Context tax shows up in three places within a workflow: discoverability (can you find what you are looking for easily?), quality (is the information you find accurate?), and workflow integrations (does up-to-date information reach you, your agent, or where you are working when you need it?).",[],{"_key":327,"_type":179,"alt":328,"asset":329},"c9c4546ac874","Top sources for work-related answers: Coworkers\u002Fteammates, code\u002Fcomments, internal docs\u002Fwikis, internal chat tools, official docs.",{"_ref":330,"_type":156},"image-526897769321e52a8823b5f33aca2a4b7aa3dbfe-883x614-png",{"_key":332,"_type":21,"children":333,"markDefs":338,"style":41},"c1acaf42e7e8",[334],{"_key":335,"_type":25,"marks":336,"text":337},"6986783f0587",[],"Consider the typical working day for a developer. They pick up a ticket that touches on a service they aren’t familiar with, so they search the company wiki and find two pieces of documentation that contradict one another. Then they find a chat from last year that references a design that is no longer current. So they ask an AI assistant and get an answer that was sourced from an outdated web page. They Slack a senior engineer to verify which piece of information is correct, but by the time they have an answer they can validate, half of their morning is gone. Even worse, none of this was captured anywhere, so another person will likely repeat this same process.",[],{"_key":340,"_type":21,"children":341,"markDefs":345,"style":41},"faf28c98b117",[342],{"_key":343,"_type":25,"marks":344,"text":48},"e1b434c2b9ab",[],[],{"_key":347,"_type":21,"children":348,"markDefs":353,"style":41},"3082f02dd1fe",[349],{"_key":350,"_type":25,"marks":351,"text":352},"6d6f8d9a8656",[],"Developers still tend to find ways around blocks to be more efficient and use AI tools they trust more than others. This further suggests that a verifiable source of knowledge is the missing layer in most workflows.",[],{"_key":355,"_type":21,"children":356,"markDefs":360,"style":41},"758723dcc682",[357],{"_key":358,"_type":25,"marks":359,"text":48},"78cba5b7905e",[],[],{"_key":362,"_type":179,"alt":363,"asset":364},"da7565cc800a","How developers manage AI context: Notes, new chat, team documentation, ask teammates, manually pick files, search multiple tools, copy\u002Fpaste between tools. ",{"_ref":365,"_type":156},"image-5e248d613477abc88adba4393bc12d17fdc6c3e5-883x630-png",{"_key":367,"_type":21,"children":368,"markDefs":372,"style":41},"42211d309eff",[369],{"_key":350,"_type":25,"marks":370,"text":371},[],"\n\nOur survey shows that whether or not organizations maintain controls of AI for their teams varies widely. 24% of respondents say their organization provides AI tools and guidance, while 11% expect tools to be used for some tasks, and 10% measure AI use as part of performance reviews. On the other hand, 30% or organizations leave AI tool decisions up to individuals, 17.3% say different teams choose their own AI tools, 32.8% say their org uses a few approved tools, and 15% encourage AI use but don’t have consistent practices.",[],{"_key":374,"_type":21,"children":375,"markDefs":379,"style":41},"c26ca6f1b001",[376],{"_key":377,"_type":25,"marks":378,"text":48},"3bbe366ba14f",[],[],{"_key":381,"_type":21,"children":382,"markDefs":387,"style":41},"b79a0579b4d9",[383],{"_key":384,"_type":25,"marks":385,"text":386},"68e4d1642e9b",[],"This shows that nearly half of developers work somewhere AI use is either left to individuals or encouraged without consistent practices in place. The best path forward to save time and standardize processes is to make sure the approved path is also the easiest to use. This means integrating shared organizational context that can be validated, kept current, and accessed by the teams and tools that need it.",[],{"_key":389,"_type":21,"children":390,"markDefs":394,"style":41},"dd64c7923e5b",[391],{"_key":392,"_type":25,"marks":393,"text":48},"b76110ab8217",[],[],{"_key":396,"_type":21,"children":397,"markDefs":402,"style":41},"29561bf11db6",[398],{"_key":399,"_type":25,"marks":400,"text":401},"992acc550a83",[],"For finance leaders, this is essential to understand: search-time savings is a leading metric that can be estimated but not easily tracked. When the time lost searching for missing context isn’t measured or at least acknowledged, it can’t be fixed. Once it is noticed, fixing it requires setting standards for what trusted context should look like, where it should be stored, and how it should be maintained going forward.",[],{"_key":404,"_type":21,"children":405,"markDefs":409,"style":41},"c0c70eeae3ff",[406],{"_key":407,"_type":25,"marks":408,"text":48},"afa4df5ce37e",[],[],{"_key":411,"_type":21,"children":412,"markDefs":417,"style":41},"41174ce1426c",[413],{"_key":414,"_type":25,"marks":415,"text":416},"69083020427d",[112],"What to do next: make checking AI faster",[],{"_key":419,"_type":21,"children":420,"markDefs":429,"style":41},"bce79906eb39",[421,425],{"_key":422,"_type":25,"marks":423,"text":424},"12cd59410784",[112],"Standardize context and use tools that integrate with what developers already use: ",{"_key":426,"_type":25,"marks":427,"text":428},"53e57aaa681c",[],"Your AI toolset should work with what your team is already utilizing. Approve access to assistants and agents that integrate well, and invest in one governed source of organizational knowledge they all draw from.",[],{"_key":431,"_type":21,"children":432,"markDefs":441,"style":41},"d6b8c98a9869",[433,437],{"_key":434,"_type":25,"marks":435,"text":436},"07d4617b9f2a",[112],"Treat pilot cancellations as a sign that you need more company-specific knowledge:",{"_key":438,"_type":25,"marks":439,"text":440},"e1b80b62214e",[]," Before cancelling a pilot over quality, ask whether the tool had access to the organizational context it needed.",[],[],{"_type":199,"current":444},"the-content-tax","The content tax",{"_key":447,"_type":160,"body":448,"sidebarCta":677,"slug":678,"title":680},"0f3f1c68679c",[449,457,464,471,476,484,491,503,515,527,539,551,558,570,582,594,606,618,625,641,653,665],{"_key":450,"_type":21,"children":451,"markDefs":456,"style":41},"94318607d912",[452],{"_key":453,"_type":25,"marks":454,"text":455},"6d9e62dd87ad",[],"79% of developers say that source attribution is important when trusting an AI answer. Further, developers tend to have confidence in knowledge sourced from colleagues the most. 72% say that coworkers are the most-used source for work related answers. ",[],{"_key":458,"_type":21,"children":459,"markDefs":463,"style":41},"0b5bb7f9c5b4",[460],{"_key":461,"_type":25,"marks":462,"text":48},"02aa0c0fe602",[],[],{"_key":465,"_type":21,"children":466,"markDefs":470,"style":41},"be909a1ab0de",[467],{"_key":453,"_type":25,"marks":468,"text":469},[],"It’s clear that established relationships and successful track records are where workers place their trust when it comes to organizational knowledge. Information sourced from AI does not get this benefit of the doubt because, understandably, developers want to know where the AI sourced the answer from before relying on it. According to our survey, the top barriers to verifiable information are Incomplete info (63%), knowledge only living in people's heads (61%), and outdated information (56%).",[],{"_key":472,"_type":179,"alt":473,"asset":474},"9e34e0bbed16","What makes information trustworthy: Up to date, from a trusted source, matches other sources, clear and complete, includes sources.",{"_ref":475,"_type":156},"image-305415eacac264c2cc810d8beb0673b84a5b4ab4-883x590-png",{"_key":477,"_type":21,"children":478,"markDefs":483,"style":41},"ae5104593037",[479],{"_key":480,"_type":25,"marks":481,"text":482},"7caff22a6e0a",[],"Five signals determine when information can be trusted:",[],{"_key":485,"_type":21,"children":486,"markDefs":490,"style":41},"5b58e112f1cc",[487],{"_key":488,"_type":25,"marks":489,"text":48},"883b91719599",[],[],{"_key":492,"_type":21,"children":493,"markDefs":502,"style":41},"07b7b0b97a32",[494,498],{"_key":495,"_type":25,"marks":496,"text":497},"7025c1b06c07",[112],"Source visibility.",{"_key":499,"_type":25,"marks":500,"text":501},"cc24e50a6f36",[]," Can I see where this came from?",[],{"_key":504,"_type":21,"children":505,"markDefs":514,"style":41},"4ff4c93b404a",[506,510],{"_key":507,"_type":25,"marks":508,"text":509},"9f8ef6fafd4b",[112],"Recency.",{"_key":511,"_type":25,"marks":512,"text":513},"0670f5dbe2ef",[]," Is it current or has it been replaced by something else?",[],{"_key":516,"_type":21,"children":517,"markDefs":526,"style":41},"4687e522485c",[518,522],{"_key":519,"_type":25,"marks":520,"text":521},"d716971b85f5",[112],"Ownership.",{"_key":523,"_type":25,"marks":524,"text":525},"82c5b57b909a",[]," Do I know who created it or who is accountable for it?",[],{"_key":528,"_type":21,"children":529,"markDefs":538,"style":41},"fb134b65d1a6",[530,534],{"_key":531,"_type":25,"marks":532,"text":533},"9e6eb8a22f13",[112],"Permissions.",{"_key":535,"_type":25,"marks":536,"text":537},"e03f45b9a52e",[]," Am I, or is this agent, allowed to see and act on it?",[],{"_key":540,"_type":21,"children":541,"markDefs":550,"style":41},"c1f1ab8d0e72",[542,546],{"_key":543,"_type":25,"marks":544,"text":545},"56f2f89c51dc",[112],"Validation.",{"_key":547,"_type":25,"marks":548,"text":549},"94e1f2343de7",[]," Has someone with the right expertise confirmed it?\n\nTo meet this standard for humans and AI agents with consistency, more than a wiki is required. Creating an organizational context that will ultimately improve the quality of AI tool outputs is done in five phases.",[],{"_key":552,"_type":21,"children":553,"markDefs":557,"style":41},"b3ef93a65b51",[554],{"_key":555,"_type":25,"marks":556,"text":48},"20f623943189",[],[],{"_key":559,"_type":21,"children":560,"level":80,"listItem":81,"markDefs":569,"style":41},"2c3f2ccdabdc",[561,565],{"_key":562,"_type":25,"marks":563,"text":564},"454c9737acde",[112],"Capture.",{"_key":566,"_type":25,"marks":567,"text":568},"d6c14f2aa174",[]," Pull knowledge from where work already happens (chat threads, tickets, wikis, repositories, and AI tools) and break it into reusable units, keeping each unit’s source attached.",[],{"_key":571,"_type":21,"children":572,"level":80,"listItem":81,"markDefs":581,"style":41},"c8532bd56820",[573,577],{"_key":574,"_type":25,"marks":575,"text":576},"8042aeb55af6",[112],"Validate.",{"_key":578,"_type":25,"marks":579,"text":580},"161f74812fe0",[]," Attach trust signals to every response, such as authorship, recency, usage, and conflict detection, and route critical gaps or conflicts to the right subject matter expert. This is where a trusted knowledge layer and verified data come in.",[],{"_key":583,"_type":21,"children":584,"level":80,"listItem":81,"markDefs":593,"style":41},"f378a125788d",[585,589],{"_key":586,"_type":25,"marks":587,"text":588},"079e214b7ead",[112],"Organize.",{"_key":590,"_type":25,"marks":591,"text":592},"a881cdc4b8a7",[]," Scope knowledge to the right projects, teams, and agents so each sees what is relevant and nothing more.",[],{"_key":595,"_type":21,"children":596,"level":80,"listItem":81,"markDefs":605,"style":41},"093a29ce4768",[597,601],{"_key":598,"_type":25,"marks":599,"text":600},"d971f8d1119e",[112],"Govern.",{"_key":602,"_type":25,"marks":603,"text":604},"7bb07cccc3d2",[]," Manage permissions for people, teams, and agents centrally, with curation activity and audit trails.",[],{"_key":607,"_type":21,"children":608,"level":80,"listItem":81,"markDefs":617,"style":41},"445e65a189d0",[609,613],{"_key":610,"_type":25,"marks":611,"text":612},"1a19acf23ec3",[112],"Deliver.",{"_key":614,"_type":25,"marks":615,"text":616},"52ce1cdee7b2",[]," Put validated knowledge into the tools developers already use, and into agents, through standard interfaces such as MCP and APIs.",[],{"_key":619,"_type":21,"children":620,"markDefs":624,"style":41},"abd681206ac3",[621],{"_key":622,"_type":25,"marks":623,"text":48},"dffff871269f",[],[],{"_key":626,"_type":21,"children":627,"markDefs":640,"style":41},"ef2fcb90eb08",[628,632,636],{"_key":629,"_type":25,"marks":630,"text":631},"ae4bd51c22cb",[],"This all matters because permissions and ownership are the biggest trust signals for practitioners, while also carrying extra weight for CISOs and compliance and risk leaders. When an agent acts on information or when an AI answer is integrated into a workflow, the team needs to be able to show where that information came from and who owns it. This will reduce the time spent verifying answers, which in turn makes teams more productive so they drive real outcomes for the business.\n\n",{"_key":633,"_type":25,"marks":634,"text":635},"0f670b27d3db",[112],"What to do next: build context, keep it current, and use it everywhere. Put search time on the budget.",{"_key":637,"_type":25,"marks":638,"text":639},"41991fa3cd8b",[]," Estimate the hours your teams spend finding information each week and track that number over time. It is the clearest business-case metric for knowledge investment.",[],{"_key":642,"_type":21,"children":643,"markDefs":652,"style":41},"39ff02be576d",[644,648],{"_key":645,"_type":25,"marks":646,"text":647},"e78f7c121be9",[112],"Capture knowledge where work happens.",{"_key":649,"_type":25,"marks":650,"text":651},"5d9f49b981a4",[]," Don’t ask developers to write documentation in yet another tool. Capture decisions and answers from the chat, tickets, and repositories they already use.",[],{"_key":654,"_type":21,"children":655,"markDefs":664,"style":41},"11af36265464",[656,660],{"_key":657,"_type":25,"marks":658,"text":659},"a454db2f32cc",[112],"Make visibility the default.",{"_key":661,"_type":25,"marks":662,"text":663},"6b9d8a2bfbeb",[]," Every answer a human or agent receives should have a clear path to track to a source, recency, owner, and validation status.",[],{"_key":666,"_type":21,"children":667,"markDefs":676,"style":41},"80803eea302b",[668,672],{"_key":669,"_type":25,"marks":670,"text":671},"b1cf79548ea4",[112],"Connect practitioners to human subject matter experts.",{"_key":673,"_type":25,"marks":674,"text":675},"82f6515e91f7",[]," Route only real gaps and conflicts to subject matter experts, then capture their answer so it is asked once, not weekly.",[],[],{"_type":199,"current":679},"the-verified-context-standard","The verified context standard",{"_key":682,"_type":160,"body":683,"sidebarCta":866,"slug":867,"title":869},"e6f3ae51662e",[684,692,699,707,714,722,729,737,744,752,760,765,773,780,788,796,803,811,818,830,842,854],{"_key":685,"_type":21,"children":686,"markDefs":691,"style":41},"e9ddd54cfa9d",[687],{"_key":688,"_type":25,"marks":689,"text":690},"fce8fe836763",[],"As AI spend keeps increasing, it’s tempting to look at it purely on the basis of cost per token or cost per seat. The real focus should be on outcomes.",[],{"_key":693,"_type":21,"children":694,"markDefs":698,"style":41},"f5333b304178",[695],{"_key":696,"_type":25,"marks":697,"text":48},"db4eaba18589",[],[],{"_key":700,"_type":21,"children":701,"markDefs":706,"style":41},"3b91c111b90c",[702],{"_key":703,"_type":25,"marks":704,"text":705},"320f2289fe2e",[],"AI is a tool your team is using to get more done faster or at least more efficiently. The results you get from using these tools are where the value lies. Without outcomes that meet your goals, the cost would be a drain on your business and cost per seat wouldn’t matter at all.",[],{"_key":708,"_type":21,"children":709,"markDefs":713,"style":41},"d7845d163a34",[710],{"_key":711,"_type":25,"marks":712,"text":48},"cba5b00da934",[],[],{"_key":715,"_type":21,"children":716,"markDefs":721,"style":41},"b4582dd4b433",[717],{"_key":718,"_type":25,"marks":719,"text":720},"dd81fa0c482a",[],"For teams that use AI, outcomes are reliant on the quality of the results they get. That quality relies entirely on improving organizational access to a knowledge base that can be relied on by every team using any AI tool.",[],{"_key":723,"_type":21,"children":724,"markDefs":728,"style":41},"8fdffc95b9d3",[725],{"_key":726,"_type":25,"marks":727,"text":48},"b2960c91ad06",[],[],{"_key":730,"_type":21,"children":731,"markDefs":736,"style":41},"3a9b3a4ec237",[732],{"_key":733,"_type":25,"marks":734,"text":735},"9495102bb2a6",[],"With this in mind, there are two important questions that finance and procurement leaders should focus on besides the cost per token:",[],{"_key":738,"_type":21,"children":739,"markDefs":743,"style":41},"133bcd29fe4b",[740],{"_key":741,"_type":25,"marks":742,"text":48},"64c9a3682f17",[],[],{"_key":745,"_type":21,"children":746,"level":80,"listItem":81,"markDefs":751,"style":41},"4fa3ba83e52f",[747],{"_key":748,"_type":25,"marks":749,"text":750},"634c5e19a6a2",[],"Does the cost change how developers are working?",[],{"_key":753,"_type":21,"children":754,"level":80,"listItem":81,"markDefs":759,"style":41},"bec5fcff1512",[755],{"_key":756,"_type":25,"marks":757,"text":758},"fd2f3f4db997",[],"What about our current workflows is holding teams back from improving AI output quality?",[],{"_key":761,"_type":179,"alt":762,"asset":763},"b771bb5f0d7e","Does cost change how developers prompt? 75% for \"Yes, to improve output quality,\" 38% for \"Yes, to reduce cost,\" and 16% for \"No.\"",{"_ref":764,"_type":156},"image-92e820cf7fe638b58c8f11b2286c103dc06f5839-883x433-png",{"_key":766,"_type":21,"children":767,"markDefs":772,"style":41},"3c15afbaccf3",[768],{"_key":769,"_type":25,"marks":770,"text":771},"2c66570d702e",[],"Most developers are working across multiple tools every week, including more than one AI tool. On top of that, the business tech stack keeps growing and changing.",[],{"_key":774,"_type":21,"children":775,"markDefs":779,"style":41},"f0daf0e3388b",[776],{"_key":777,"_type":25,"marks":778,"text":48},"32c6b8cc60d6",[],[],{"_key":781,"_type":21,"children":782,"markDefs":787,"style":41},"3920a546bddc",[783],{"_key":784,"_type":25,"marks":785,"text":786},"dc7e1004cd9a",[],"Our survey shows that 19.7% of practitioners are using 1 tool, while the majority at nearly 75% are using between two and five different AI tools.",[],{"_key":789,"_type":21,"children":790,"markDefs":795,"style":41},"f1f87cd2c088",[791],{"_key":792,"_type":25,"marks":793,"text":794},"6c41aebdd435",[],"Teams need room to evolve their tech stack as needs and goals shift. When all of your team’s tools are sourcing answers from the same validated, permission-aware source of information, the specific AI tools used matters less. With context in place your team can use the AI they prefer and new tools are more easily adopted as needed.",[],{"_key":797,"_type":21,"children":798,"markDefs":802,"style":41},"7cb406a25f0e",[799],{"_key":800,"_type":25,"marks":801,"text":48},"5f69a7021a8c",[],[],{"_key":804,"_type":21,"children":805,"markDefs":810,"style":41},"171032432ea9",[806],{"_key":807,"_type":25,"marks":808,"text":809},"2dab68cb8b52",[],"The goal is ensuring that any technology you and your team use is context-aware enough to need fewer corrections because it has been provided with validated and current information from the start.",[],{"_key":812,"_type":21,"children":813,"markDefs":817,"style":41},"6b93756949cb",[814],{"_key":815,"_type":25,"marks":816,"text":48},"6c5cce8d9d40",[],[],{"_key":819,"_type":21,"children":820,"markDefs":829,"style":41},"265a71c7340d",[821,825],{"_key":822,"_type":25,"marks":823,"text":824},"fb60c2d8d3f6",[112],"What to do next: focus on outcomes, not tokens. Retire usage as a success metric.",{"_key":826,"_type":25,"marks":827,"text":828},"70d840f619ff",[]," Tokens consumed, prompts sent, and seats activated measure activity. Set goals on cycle time, correction and rework rates, search time saved, and incidents avoided.",[],{"_key":831,"_type":21,"children":832,"markDefs":841,"style":41},"62e769518f28",[833,837],{"_key":834,"_type":25,"marks":835,"text":836},"69e7b6a2beba",[112],"Don’t tie performance reviews to AI usage.",{"_key":838,"_type":25,"marks":839,"text":840},"80c27c537aa1",[]," Measuring individuals on how much AI they use rewards volume over value and pushes developers to use AI where it doesn’t help. Measure team outcomes instead.",[],{"_key":843,"_type":21,"children":844,"markDefs":853,"style":41},"14352f7b8932",[845,849],{"_key":846,"_type":25,"marks":847,"text":848},"b0de6218c4ef",[112],"Forecast spend against outcomes.",{"_key":850,"_type":25,"marks":851,"text":852},"1ffcaff7733e",[]," Pair usage limits with outcome reporting, so spend can increase where results justify it rather than being capped across the board.",[],{"_key":855,"_type":21,"children":856,"markDefs":865,"style":41},"5f3bf5a94df9",[857,861],{"_key":858,"_type":25,"marks":859,"text":860},"7ff19607cce4",[112],"Standardize the context layer.",{"_key":862,"_type":25,"marks":863,"text":864},"51bb76ddfd2c",[]," Keep the assistant and agent layer flexible, and make the shared knowledge layer the stable, governed foundation.",[],[],{"_type":199,"current":868},"why-the-outcomes-drive-the-real-value-of-ai","Why the outcomes drive the real value of AI",{"_key":871,"_type":160,"body":872,"sidebarCta":977,"slug":978,"title":980},"a0be84f34a55",[873,881,888,896,903,911,918,930,942,954,962],{"_key":874,"_type":21,"children":875,"markDefs":880,"style":41},"5b21c1db8c70",[876],{"_key":877,"_type":25,"marks":878,"text":879},"7f72c59d2a9d",[],"Reliability, context, and cost each route to three different leadership seats looking at AI from a buyer perspective. The CIO sees pilots that stall. The CTO and platform leaders see teams losing hours to search and rework each week. The CFO and procurement leaders see spend that is hard to tie to results.",[],{"_key":882,"_type":21,"children":883,"markDefs":887,"style":41},"1f0ef019071c",[884],{"_key":885,"_type":25,"marks":886,"text":48},"03bebaa0e6de",[],[],{"_key":889,"_type":21,"children":890,"markDefs":895,"style":41},"0aafda7acb52",[891],{"_key":892,"_type":25,"marks":893,"text":894},"3dbd5940a953",[],"All three need a foundation of reliable context to ensure AI is driving outcomes and not more problems.",[],{"_key":897,"_type":21,"children":898,"markDefs":902,"style":41},"1d3b18c718cd",[899],{"_key":900,"_type":25,"marks":901,"text":48},"353ad8b4c2db",[],[],{"_key":904,"_type":21,"children":905,"markDefs":910,"style":41},"6e2beb2c0648",[906],{"_key":907,"_type":25,"marks":908,"text":909},"ca18f80d71e6",[],"The solution is implementing a verifiable knowledge layer:",[],{"_key":912,"_type":21,"children":913,"markDefs":917,"style":41},"5bbd8837e1fd",[914],{"_key":915,"_type":25,"marks":916,"text":48},"3fb9a6f52234",[],[],{"_key":919,"_type":21,"children":920,"level":80,"listItem":81,"markDefs":929,"style":41},"5f6a8b8b8b89",[921,925],{"_key":922,"_type":25,"marks":923,"text":924},"71098b3f1939",[112],"Verification built in.",{"_key":926,"_type":25,"marks":927,"text":928},"029ad5eb5fa8",[]," Output arrives grounded in validated knowledge, so checking it takes seconds, not hours.",[],{"_key":931,"_type":21,"children":932,"level":80,"listItem":81,"markDefs":941,"style":41},"18ffcc3ca82c",[933,937],{"_key":934,"_type":25,"marks":935,"text":936},"f15d76c3c9bd",[112],"Established knowledge.",{"_key":938,"_type":25,"marks":939,"text":940},"07a13c73e1b0",[]," Every answer, for every human and agent, carries its source, recency, owner, permissions, and validation status.",[],{"_key":943,"_type":21,"children":944,"level":80,"listItem":81,"markDefs":953,"style":41},"ce2420b3b6be",[945,949],{"_key":946,"_type":25,"marks":947,"text":948},"4339e124898f",[112],"Outcome-based measurement.",{"_key":950,"_type":25,"marks":951,"text":952},"d80aafa2fa22",[]," Success is measured in better outcomes for the actual business, less search time, and predictable spend, not tokens or seats.",[],{"_key":955,"_type":21,"children":956,"markDefs":961,"style":41},"a04b3c8c1788",[957],{"_key":958,"_type":25,"marks":959,"text":960},"e2c5488cf5db",[],"Stack Internal was built to solve these challenges as a trusted knowledge layer that synthesizes and verifies organizational context across your business. This knowledge is then delivered to teams and into the tools they already use, through integrations such as Slack, Microsoft Teams, Jira, Confluence, and GitHub, and to agents through MCP and API.",[],{"_key":963,"_type":21,"children":964,"markDefs":974,"style":41},"42ccad875314",[965,970],{"_key":966,"_type":25,"marks":967,"text":969},"e93aff333fcc",[968],"f708d40d568c","Talk with an expert",{"_key":971,"_type":25,"marks":972,"text":973},"f2dd44734c02",[]," today to learn more about how adding a knowledge layer to your organization would improve AI workflows and drive better outcomes.",[975],{"_key":968,"_type":39,"href":976},"https:\u002F\u002Fstackoverflow.co\u002Finternal\u002F",[],{"_type":199,"current":979},"valuable-ai-outcomes-start-with-the-right-context","Valuable AI outcomes start with the right context",true,{"_type":179,"alt":983,"asset":984},"Stack Overflow graphic featuring the text: \"Why Your AI Almost Works: What the 2026 Developer Survey Says About Business Context.\"",{"_ref":985,"_type":156},"image-d2bb08a529b83c49eb83bcd298c014fbc2ccbbef-1200x630-png",{"_createdAt":987,"_id":988,"_rev":989,"_system":990,"_type":993,"_updatedAt":994,"abbr":995,"color":996,"descriptionShort":1013,"features":1014,"inflection":1030,"marketo":1030,"name":1031,"nameFull":1032,"plans":1033,"slug":1040},"2022-02-25T09:59:05Z","339bc91a-69c6-4a69-8add-a12977a22ad5","aUGu5vJZMW8mvj74UZYZiv",{"base":991},{"id":988,"rev":992},"PhasGW24ij3IO6Qup2krVc","product","2026-09-30T15:45:20Z","SOI",{"_type":997,"alpha":80,"hex":998,"hsl":999,"hsv":1004,"rgb":1008},"color","#2b2d6e",{"_type":1000,"a":80,"h":1001,"l":1002,"s":1003},"hslaColor",238.2089552238806,0.30000000000000004,0.4379084967320261,{"_type":1005,"a":80,"h":1001,"s":1006,"v":1007},"hsvaColor",0.6090909090909091,0.43137254901960786,{"_type":1009,"a":80,"b":1010,"g":1011,"r":1012},"rgbaColor",110,45,43,"Where developers & technologists share private knowledge with coworkers.",[1015,1018,1021,1024,1027],{"_key":1016,"_ref":1017,"_type":156},"1d36a6b63dc4","7c088242-c7bb-4d91-93db-6f7a042b1484",{"_key":1019,"_ref":1020,"_type":156},"f335b465079d","productFeatureCategory-integrations",{"_key":1022,"_ref":1023,"_type":156},"47ec4cc75ac7","productFeatureCategory-admin-support",{"_key":1025,"_ref":1026,"_type":156},"0c915c1f0429","productFeatureCategory-security",{"_key":1028,"_ref":1029,"_type":156},"e3e4701a58d6","productFeatureCategory-community","Stack Internal","Internal","Stack Overflow Internal",[1034,1037],{"_key":1035,"_ref":1036,"_type":156},"5d394ab03849","090566fe-d9d8-4014-bd3d-725502428be6",{"_key":1038,"_ref":1039,"_type":156},"ec5f82035b82","95303e50-2a79-4bcf-b317-f0fc0e44a8ef",{"_type":199,"current":1041},"internal","2027-11-01T15:57:00.000Z",[1044],{"_key":1045,"_ref":1046,"_type":156},"302eb9f2f81a","b4721a78-6640-4834-b1fd-d2f6e434a4f4",{"_type":199,"current":1048},"2026-developer-survey-for-business-leaders",{"_ref":1050,"_type":156},"c18a6a4e-32f6-4b21-ad23-67d6add15691","What the 2026 Developer Survey says about business context"]