Knowledge Management Statistics (2026)

A data analysis of how much time teams lose hunting for answers, what walks out the door when people quit, why customers give up on help centers, and whether AI search actually fixes any of it.

The headline knowledge management statistic for 2026: teams spend about 25% of their working time just searching for answers, according to Atlassian’s State of Teams 2025 survey of 12,000 knowledge workers. A 2022 Glean and Harris Poll survey landed on the same quarter of the workweek. The rest of the data points one way: company knowledge is scattered, thinly documented, and leaves with the people who hold it. That is the case for one well-kept knowledge base over several half-maintained wikis.

Key takeaways

• 25% of working time goes to searching for answers, per Atlassian (2025); Glean’s 2022 Harris Poll survey found the same 25% of the workweek
• 2.4 billion hours a year are lost to searching for information across the Fortune 500, per Atlassian
• 42% of institutional knowledge is unique to the person who holds it, per a 2018 Panopto study reported by HR Dive, so coworkers can’t do 42% of that job once the person leaves
• 56% of employees still have to ping someone or schedule a meeting to get information they need (Atlassian, 2025)
• 43% of failed self-service attempts happen because customers can’t find content relevant to their issue, per Gartner (2024)
• 51% of customer service journeys now start on third-party platforms such as Google, YouTube, and ChatGPT (Gartner, 2025)
• 61% of developers spend more than 30 minutes a day searching for answers, per the 2024 Stack Overflow Developer Survey; in 2025, finding information led developers’ list of top time-wasters (Atlassian)
• 53% of workers say important information they need is not accessible from their AI tools, per Glean’s Work AI Index 2026
• The knowledge management software market is sized at $13.70B to $23.2B for 2025, per Mordor Intelligence and Fortune Business Insights

How much time do employees spend searching for information?

Roughly a quarter of the workweek. Atlassian’s State of Teams 2025 found leaders and teams waste 25% of their time just searching for answers, and a 2022 Glean and Harris Poll survey reached the same 25%. Coveo’s 2025 survey puts it at three hours a day.

1. Teams waste 25% of their time just searching for answers, per Atlassian’s survey of 12,000 knowledge workers and 200 executives. That is more than a full day every week spent locating work instead of doing it.
2. Fortune 500 companies collectively waste 2.4 billion hours a year searching for information (Atlassian). Atlassian blames a habit of treating file storage as if it were knowledge management.
3. Employees waste an average of three hours per day searching for information, per Coveo’s 2025 EX Relevance Report, a survey of 4,000 US and UK workers at companies with 5,000+ employees. They also say 42% of the information they sift through is irrelevant to their role. Coveo sells AI search, so read this as vendor research.
4. Employees spend at least two hours a day, or 25% of the workweek, looking for documents, information, or people, per Glean’s 2022 survey of 1,043 knowledge workers, run by The Harris Poll. Glean also sells enterprise search.
5. 62% of workers say they struggle with too much time spent searching for information, per Microsoft’s 2023 Work Trend Index, which surveyed 31,000 people across 31 markets.
6. The average interaction worker spent nearly 20% of the workweek looking for internal information or tracking down colleagues, per a 2012 McKinsey Global Institute analysis. The same report estimated a searchable record of knowledge can cut search time by as much as 35%.

Sources: Atlassian, Glean, Coveo, Microsoft, Stack Overflow, McKinsey. Definitions and samples differ, so compare the range, not the decimals.

Across more than a decade of studies by different firms, the estimate barely moves: a fifth to a quarter of knowledge work goes to finding things. Our note-taking statistics track Coveo’s earlier daily search figures, and our productivity software statistics cover the related cost of app switching.

What does poor knowledge sharing cost a business?

Mostly duplicated effort and waiting. A 2018 Panopto study, reported by HR Dive, estimated the average large US business loses $47 million in productivity a year to inefficient knowledge sharing. In 2025, Atlassian found half of employees say teams duplicate work without realizing it.

7. Large US businesses lose $47 million in productivity a year to inefficient knowledge sharing, per the 2018 Panopto Workplace Knowledge and Productivity Report (HR Dive). Panopto tied the loss to employees waiting on colleagues and recreating information that already exists.
8. Knowledge workers waste 5.3 hours every week waiting for information from coworkers or recreating existing knowledge (Panopto via HR Dive, 2018). That is most of a workday, every week, per person.
9. 56% of employees still have to ping someone or schedule a meeting to get the information they need (Atlassian, 2025). Each ping also interrupts the person who knows the answer, so the cost lands twice.
10. Half of employees say teams at their company are duplicating work without realizing it (Atlassian). Duplicate work is the most expensive symptom, because nobody sees it on a budget line.
11. 74% of leaders say poor communication is directly hurting their team’s speed and output (Atlassian).
12. Atlassian puts the coordination “fragmentation tax” on the Fortune 500 at $161 billion a year, and 87% of knowledge workers say they lack the time or capacity to coordinate, per its State of Teams 2026. Atlassian sells Confluence and Jira, so treat its numbers as vendor research.

The dollar figures come from vendors with a product to sell, and the Panopto estimate is now eight years old. The pattern is still consistent: the cost is not one big failure but thousands of small waits, re-asks, and rebuilds.

How much knowledge is lost when employees leave?

Close to half of a role. Panopto’s 2018 study found 42% of institutional knowledge is unique to the individual, so coworkers cannot perform 42% of that job after a departure (HR Dive). With median US job tenure at 4.1 years (BLS, 2026), that loss arrives on a predictable schedule.

13. 42% of institutional knowledge is unique to the employee in their role and not shared with coworkers (Panopto via HR Dive, 2018). When that person leaves, the knowledge leaves too.
14. 65% of employees at high-turnover companies find it difficult to nearly impossible to get the information they need (Panopto via HR Dive, 2018). Turnover and poor documentation compound each other.
15. Median employee tenure in the US was 4.1 years in January 2026, up from 3.9 years in January 2024, per the Bureau of Labor Statistics. Median tenure counts years in the current job so far, and it is short next to the life of most internal systems and processes.
16. Workers aged 55 to 64 have a median tenure of 9.6 years, versus 3.0 years for ages 25 to 34 (BLS). Your longest-tenured people hold the deepest undocumented knowledge, and they are the closest to retirement.
17. US employers recorded 3.1 million quits in August 2026 alone, a quits rate of 1.9%, per BLS JOLTS data. Total separations were 5.1 million for the month.
18. 43% of employees would consider leaving a job if there wasn’t an efficient way to access the information and people they need (Glean / Harris Poll, 2022). Poor knowledge access is itself a reason people quit.

Do customers prefer self-service knowledge bases?

Yes, most try self-service first. Gartner found 73% of customers use self-service at some point in their service journey, and 2017 research published in Harvard Business Review found 81% attempt to resolve issues themselves before calling a live rep. Demand is not the problem. Resolution is.

19. 73% of customers use self-service at some point in their customer service journey, per a Gartner survey of 5,728 customers conducted in December 2023.
20. 81% of customers attempt to take care of matters themselves before reaching out to a live representative, per 2017 CEB research published in Harvard Business Review.
21. 51% of customer service journeys now begin on third-party platforms such as Google, YouTube, and ChatGPT, per a Gartner survey of 5,801 customers in early 2025. Your help center competes with search engines and chatbots for the first click.
22. 74% of Gen Z customers start their service journey with a third party before reaching an official company channel (Gartner, 2025).
23. 60% of customer service agents fail to promote self-service options (Gartner, 2025). When agents do mention it, 25% make neutral comments and 12% make explicitly negative remarks.
24. 69% of customers want AI-powered resolutions for complex service issues, and 48% want educational tools or guides, per Coveo’s 2025 CX Relevance Report of 4,000 US and UK customers.

Why do customer knowledge bases fail?

Mostly because people cannot find the right article. In Gartner’s 2024 research, the most common reason self-service failed was that in 43% of cases customers couldn’t find content relevant to their issue. Only 14% of issues were fully resolved in self-service, a gap our customer support statistics cover in depth.

25. Only 36% of even “very simple” issues resolve fully in self-service (Gartner). If the easy questions fail, the content is the problem, not the customer.
26. 45% of customers who started in self-service said the company didn’t understand what they were trying to do (Gartner). Articles get written in the company’s vocabulary, then searched in the customer’s.
27. Customers report a 62% success rate on third-party platforms like YouTube or Google, but only 22% start, stay, and resolve entirely in a company’s own channels (Gartner, 2025). Public answers on other sites are beating many official help centers.
28. 84% of customers struggle to find information, and 53% cite website search as their biggest self-service frustration (Coveo, 2025).
29. 72% of customers abandon websites after negative experiences, though 62% are willing to try again (Coveo, 2025).

Sources: Gartner 2024, Gartner 2025, Harvard Business Review, Coveo. Gartner pages were retrieved from Internet Archive copies of gartner.com.

How do developers find answers and use documentation?

Developers rely on documentation more than any other resource, yet still lose time hunting. In the 2025 Stack Overflow survey, 67.8% used technical documentation to learn. In 2024, 61% spent over 30 minutes a day searching for answers (Stack Overflow), and in 2025 finding information led developers’ list of time-wasters (Atlassian).

30. 61% of developers spend more than 30 minutes a day searching for answers or solutions to problems (Stack Overflow 2024, 28,911 responses). That count includes asking a colleague and waiting for a reply.
31. 53% of developers agree that waiting on answers disrupts their workflow, even when they know where to look (Stack Overflow 2024). 70% know where to go for answers, but only 56% say they can find them quickly.
32. 30% of developers say knowledge silos hit their productivity ten or more times a week (Stack Overflow 2024), and 45% agree that silos are a nuisance.
33. 50% of developers lose 10+ hours a week to organizational inefficiencies, and 90% lose 6+ hours (Atlassian State of DevEx 2025, 3,500 developers and managers). Finding information (services, docs, APIs) now heads the list of top time-wasters.
34. 63% of developers say leaders don’t understand their pain points, up from 44% the year before (Atlassian). Developers spend only 16% of their time coding, so friction outside the editor dominates.
35. 67.8% of developers used technical documentation to learn to code in the past year, the most used resource, ahead of Stack Overflow at 51.4% and AI tools at 44% (Stack Overflow 2025).
36. 93% of API teams face collaboration blockers, and Postman names documentation scattered across too many places as the root cause (Postman State of the API 2025, more than 5,700 respondents). 41% of respondents already rely on AI to generate API documentation.

The developer data mirrors the broader workforce: most people know where to look, yet answers still arrive slowly. Postman’s diagnosis applies far beyond APIs. Teams rarely lack documentation; they have too many copies of it in too many tools.

Is AI search fixing how teams find company knowledge?

Not yet. AI use is close to universal, but the knowledge underneath is the bottleneck. Glean’s Work AI Index 2026 found 87% of digital workers use AI at work, yet 53% say important information they need is not accessible from their AI tools, and only 13% say AI significantly improved their organization’s performance.

37. 85% of knowledge workers use AI at work, but only 29% have embedded it in their flow of work (Atlassian State of Teams 2026, 12,035 knowledge workers). Usage is wide and shallow.
38. 96% of executives aren’t sure how to get their teams to use AI effectively (Atlassian, 2025). AI can only work with knowledge it can see.
39. 53% of workers say important information they need is not accessible from their AI tools (Glean, survey of 6,000 workers, 2026). Workers on context-poor AI are far more likely to feel worn out by it: 50% versus 18%.
40. Workers spend 37% of their AI time “botsitting”, feeding context, supervising outputs, and cleaning up, versus 36% actually using AI to do work (Glean, 2026). Glean sells enterprise AI search, so weigh it as vendor research.
41. 49% of employees have experienced AI hallucinations at work (Coveo EX 2025). On the customer side, 49% have seen them too and 42% fact-check generated answers (Coveo CX 2025).
42. 46% of developers distrust the accuracy of AI tool output, versus 33% who trust it (Stack Overflow 2025). 66% cite “AI solutions that are almost right, but not quite,” and 75% would still ask a person when they don’t trust AI’s answers.
43. 86% of workers want AI to help them find the right information and answers they need (Microsoft Work Trend Index 2023).
44. Knowledge management is a top-four generative AI use case at 28% of respondents, behind customer self-service at 34% (Coveo EX 2025).

The pattern repeats across every source: AI search is only as good as the knowledge it can reach and the permissions it respects. A tidy, current knowledge base is the input that makes AI answers trustworthy.

Why do more tools make knowledge harder to find?

Every new app is another place to search. Gartner’s 2023 survey found 47% of digital workers struggle to find the information they need, while the average desk worker now uses 11 applications, up from six in 2019. Coveo’s 2025 survey also landed on 47%: that share say the information they need is stored across multiple applications.

45. 47% of digital workers struggle to find the information or data needed to perform their jobs (Gartner, survey of 4,861 employees run September through November 2022, published 2023).
46. The average desk worker uses 11 applications, compared with six in 2019; 40% use more than the average and 5% use 26 or more (Gartner, 2023).
47. 66% of digital workers agree better business outcomes would follow if IT provided universally accepted applications (Gartner, 2023). Workers are asking for fewer, standard tools.
48. 47% of employees say the information they need is stored across multiple applications, the largest hurdle year over year (Coveo, 2025).
49. Employees search four different data sources on average to find information, and 25% don’t know where to begin (Coveo, 2025).
50. 77% of AI users juggle multiple AI tools each week, and 33% use four or more (Glean, 2026). AI tool sprawl is repeating the wiki sprawl pattern one layer up.

This is the core buying lesson. A second wiki does not halve the search problem; it doubles the number of places an answer might be stale. Teams that standardize on one Confluence or Notion workspace, and keep it current, give both people and AI a single place to look. Lighter options such as Coda for doc-plus-data work, or Obsidian for personal notes that live alongside the team wiki (see the note-taking category), work when their role is clearly bounded.

How big is the knowledge management software market?

Estimates for 2025 range from $13.70 billion to $23.2 billion. Mordor Intelligence projects $37.64 billion by 2031 at an 18.34% CAGR, while Fortune Business Insights projects $74.22 billion by 2034 at 13.8%. The adjacent enterprise search market is $7.47 billion in 2026, per Mordor.

51. Mordor Intelligence sizes knowledge management software at $13.70 billion in 2025, $16.22 billion in 2026, and $37.64 billion by 2031, an 18.34% CAGR (Mordor Intelligence). It credits the growth to generative AI tools that capture and act on company know-how.
52. Fortune Business Insights sizes the same market at $23.2 billion in 2025 and $26.4 billion in 2026, reaching $74.22 billion by 2034 at a 13.8% CAGR (Fortune Business Insights). Research firms draw the market’s boundaries differently, so read the two estimates as a range.
53. The enterprise search market reached $7.47 billion in 2026 and is projected to hit $11.66 billion by 2031, a 9.31% CAGR (Mordor Intelligence). Mordor ties the growth to retrieval-augmented generation and conversational interfaces replacing keyword search.

Sources: Mordor Intelligence, knowledge management, Fortune Business Insights, Mordor Intelligence, enterprise search. Market sizes are research-firm estimates and vary with scope.

What to do with this data

The numbers describe one problem from several angles: knowledge spread across too many tools, written once and rarely maintained, and held in a few heads. If you are choosing a knowledge base, these are the moves the data supports.

• Pick one source of truth and retire the rest. Workers already search four sources on average (Coveo) across 11 apps (Gartner). Compare the scored picks in our knowledge base category and commit to one.
• Document the 42% first. Start with your longest-tenured people, whose median tenure is 9.6 years for ages 55 to 64 (BLS). Exit interviews are too late to capture what only they know.
• Measure findability, not page count. 43% of self-service failures happen because customers can’t find relevant content (Gartner). Track failed searches and fix the top ones every month.
• Publish customer answers where customers look. With 51% of service journeys starting on Google, YouTube, or ChatGPT (Gartner), a public, indexable help center beats a gated one.
• Fix the content before buying AI search. 53% of workers say their AI tools can’t reach what they need (Glean). Choose a knowledge base whose built-in AI answers from your own pages and respects permissions.
• Give one owner the job. If operations owns the wiki at your company, our guide for operations managers covers the tools that make that role manageable.

For companion data, see our note-taking statistics, customer support statistics, and productivity software statistics. For how we evaluate and score tools, see how tools8020 works.

Frequently asked questions

What are the most important knowledge management statistics?

Three numbers frame the topic. Teams spend about 25% of their time searching for answers (Atlassian, 2025). 42% of a role’s knowledge is unique to one person (Panopto via HR Dive, 2018). And 43% of failed self-service attempts happen because customers couldn’t find relevant content (Gartner, 2024).

How much of the workweek goes to searching for information?

Estimates cluster around a fifth to a quarter of the week. Atlassian puts it at 25% of working time in 2025. Coveo measured three hours a day in 2025, and a 2012 McKinsey analysis found nearly 20% of the workweek.

What percentage of knowledge is lost when an employee leaves?

About 42%. Panopto’s 2018 study found 42% of institutional knowledge is unique to the individual in their role, so coworkers can’t perform that share of the job after a departure (HR Dive). With 3.1 million quits in the US in August 2026 alone (BLS), the loss is constant.

Do customers prefer self-service over talking to an agent?

Most try it first. Gartner found 73% of customers use self-service at some point, and 2017 research in HBR found 81% try to solve issues themselves before contacting a live rep. Only 36% of very simple issues fully resolve there, so preference outruns results.

Why do knowledge base articles fail customers?

Customers can’t find the right content. Gartner traced 43% of self-service failures to missing relevant content, and 45% of customers said the company didn’t understand what they were trying to do. Coveo found 53% name site search as their top frustration.

Does AI search solve knowledge management problems?

Only when the underlying knowledge is reachable and current. Glean’s 2026 index found 87% of workers use AI, but 53% say needed information isn’t accessible to their AI tools. Stack Overflow found 46% of developers distrust AI output accuracy. Clean, current source content is what makes AI answers worth trusting.

How big is the knowledge management software market in 2026?

Between $16.22 billion and $26.4 billion, depending on the firm. Mordor Intelligence projects $37.64 billion by 2031, and Fortune Business Insights projects $74.22 billion by 2034. The enterprise search market adds $7.47 billion in 2026 (Mordor). Treat all three as research-firm estimates with different scopes.

How we compiled these statistics

We drew every figure on this page from 25 sources, and each one was retrieved from the publisher’s page in October 2026; Gartner and McKinsey pages were read through Internet Archive copies of their own sites because the live pages block automated access. Figures we could not trace to an original publisher were omitted, and vendor research from Atlassian, Coveo, and Glean is labeled as such. Panopto’s original 2018 report page is offline, so its figures are cited through HR Dive’s coverage from the time. Where sources disagree, as on search time and market size, we show the range and cite both.

Sources

• Atlassian: State of Teams 2025 (2025)
• Atlassian: Stop mistaking storage for strategy (2025)
• Atlassian: The State of Teams 2026 (2026)
• Atlassian: State of Developer Experience 2025 (2025)
• Stack Overflow: 2024 Developer Survey, Professional Developers (2024)
• Stack Overflow: 2025 Developer Survey, AI (2025)
• Stack Overflow: 2025 Developer Survey, Developers (2025)
• HR Dive: Inefficient knowledge-sharing costs large US businesses $47M a year, on Panopto’s report (2018)
• Gartner: Only 14% of Customer Service Issues Are Fully Resolved in Self-Service (2024, Internet Archive copy)
• Gartner: More Than Half of Customer Service Journeys Now Start on Third-Party Platforms (2025, Internet Archive copy)
• Gartner: 60% of Customer Service Agents Fail to Promote Self-Service (2025, Internet Archive copy)
• Gartner: 47% of Digital Workers Struggle to Find the Information Needed to Perform Their Jobs (2023, Internet Archive copy)
• Harvard Business Review: Kick-Ass Customer Service (2017)
• Coveo: EX Relevance Report Reveals 42% of Information Fails Employees (2025)
• Coveo: CX Relevance Report Reveals 84% of Customers Struggle to Get Relevant Help (2025)
• Microsoft: Work Trend Index, Will AI Fix Work? (2023)
• Glean: Work AI Index 2026 (2026)
• Glean: Hybrid Workplace Habits and Hangups Report (2022)
• McKinsey Global Institute: The social economy (2012, Internet Archive copy)
• US Bureau of Labor Statistics: Employee Tenure Summary (2026)
• US Bureau of Labor Statistics: Job Openings and Labor Turnover Summary (2026)
• Postman: 2025 State of the API Report (2025)
• Mordor Intelligence: Knowledge Management Software Market (2026)
• Fortune Business Insights: Knowledge Management Software Market (2026)
• Mordor Intelligence: Enterprise Search Market (2026)