Feature
White Paper
The Future of Leadership Development: AI-Driven Manager Enablement
In Brief
- Wikis are making a comeback. Once dismissed as a 2000s-era relic, they're back in demand as businesses look for reliable, human-curated content to ground AI.
- Modern wikis often don't call themselves wikis. Notion, Confluence, SharePoint and Teams now do the job — collaborative, interlinked pages where teams build knowledge together.
- AI fixes the wiki's biggest weakness. Manual upkeep was always the Achilles' heel. AI agents now flag stale pages, cluster related content and keep wikis current.
When most people hear the word “wiki,” they think of Wikipedia. But wikis have been part of the digital workplace toolset for years, used for knowledge-sharing and collaboration. And while the “classic” days of the wiki as a knowledge management (KM) and enterprise collaboration tool may have passed, wikis are making a comeback as businesses scramble to feed AI reliable context.
What Is a Wiki?
Microsoft defines a wiki as “a site that is designed for groups of people to quickly capture and share ideas by creating simple pages and linking them together.”
A wiki is a collaborative site, typically accessed through a browser, where any and all members of the team contribute and edit content and knowledge. It removes barriers to editing, so wiki editing is usually done within the browser without special editing tools.
The wiki’s power — and to some critics, its downfall — comes from its community-driven or crowdsourced model. Anyone, with the right permissions, can contribute content to the wiki, meaning the wiki’s development is driven by the user community rather than individuals or a company. It is bottom-up rather than top-down.
Where Wikis Live Now
In the 2000s and 2010s, enterprise wikis were frequently a feature of classic KM initiatives, such as at investment bank Dresdner Kleinwort Wasserstein, but does the heyday of enterprise wikis lie in the past?
“While organizations may not always refer to them as ‘wikis’ anymore, they continue to use them because they solve a fundamental challenge: capturing and transferring knowledge before it is lost,” said Lynda Braksiek, principal research lead, knowledge management at APQC.
Wikis provide a collaborative environment where knowledge is continuously refined rather than left as static documentation, Braksiek said, such as standard operating procedures, project lessons learned, troubleshooting guides and capturing expert know-how.
“Wikis are still being used internally,” said Ilana Botha, an independent KM consultant at Opus Excellence. “But rather than being used organization-wide, I am seeing wikis being used in smaller, localized cases.”
They remain popular among Communities of Practice and interest groups, while also at times being used at a departmental and team level, or in programmatic and project work to onboard new members, Botha added. “This indicates a shift towards domain-specific knowledge clusters, moving away from monolithic, unmanaged corporate wikis,” she said.
The Wiki Software Landscape, From Confluence to Notion
The wiki software landscape has evolved. Enterprise wiki software is still commercially available including Mediawiki, although some earlier products such as Socialtext have been retired. Microsoft Wiki features previously available through Teams have been replaced by OneNote. Confluence, still regarded as classic KM software, is favored by IT functions, often in conjunction with other Atlassian products.
However, other collaborative solutions are also used as a wiki, even though those tools might not be recognized specifically as classic wiki software.
“Increasingly, organizations are managing knowledge in platforms such as SharePoint, Microsoft Teams, Confluence, Notion and other collaborative workspaces,” said Braksiek. If employees collaboratively create, update, organize and discover knowledge dynamically, then the platform is serving many of the same purposes as a traditional wiki, she added.
Today’s wiki use is reflected in team collaboration tools such as Notion and Slite, Botha agreed. “To an extent, modern work management tools can be seen as the evolution of the wiki,” she said. “Many have adopted the architectural principles of a wiki, including interlinked pages and collaborative editing, without using the 'wiki' label.” Teams use tools such as Notion to build dynamic knowledge bases where databases and project tracking are stored alongside standard text documents, she added.
Why Wikis Still Earn Their Keep
Wikis remain a good platform for user-generated knowledge sharing, which in turn provides authoritative content in the age of AI, Botha said. “One of the greatest advantages of wikis is that they capture tacit knowledge such as context, decisions and expertise that formal documents do not,” she said. Busy, community-driven wikis provide rich internal context that supports AI readiness, providing valuable context that AI systems need to generate reliable answers.
Knowledge transfer is most effective when documentation captures not only what to do, but also why decisions were made, common pitfalls and lessons learned, Braksiek agreed. “Wikis support this because they allow knowledge to evolve through contributions from multiple experts rather than relying on a single author or point-in-time document,” she said.
Another advantage of wikis is that they are relatively low-cost and low-effort to set up and run, and should, at least theoretically, then largely look after themselves.
The power of using a wiki in a tool like Notion also lets you embed tables or integrated data from other sources, as well as other integrated features that are dynamically updated, to provide context to knowledge or content in the wiki and vice versa. Comments and discussions also provide important context for entries on wiki topics.
“The highest-value wikis today don't look like the static repositories many people think of from 10 or 15 years ago,” said Braksiek, One APQC member company transformed its enterprise wiki into an AI-enhanced knowledge platform that supports thousands of active pages and daily users, she said.
What Wikis Go Wrong
The “anyone can edit” aspect of wikis is often held up as a barrier to their use due to the misconception that they will be inherently inaccurate. However, while this criticism is valid, particularly as content might go out of date, a successful wiki will be self-policing; everything is open to review by the whole community, allowing subject matter experts to comment, challenge, change and contribute quickly and easily.
To make this easier, changes should be also visible and trackable within wiki software, so they can be rolled back if needed. This reduces duplication of effort and means the information presented is more likely to be timely and accurate.
Another problem cited is that wikis can turn into a wild west, where knowledge workers run rampant with little to no supervision. Correspondingly, the opposite can also be true where wikis don’t get populated, turning into ghost towns.
The solution to both these issues is to have a steward for the wiki, encouraging participation and applying some light governance to maintain standards. Realistically, though, this involves administrative overhead, which undermines the purpose of the wiki as a low-cost, high-value format.
Engaged, But Not Convinced: What's Eroding Beneath Your Strongest Engagement Numbers
Watch Now
Retention Is a Communication Problem: New Data from 6,000+ Employees
Watch Now
The Wiki, Rewired for AI
With wikis a mature solution, are they now “old school” or even obsolete in the age of AI? Far from it.
“Generative AI changes the experience of using a wiki more than it changes the need for having one,” said Braksiek. “AI can dramatically improve how knowledge is captured, summarized, and discovered, but it still depends on accurate, governed content.”
“A lifecycle agent could be used to monitor the content lifecycle and ensure content is reviewed, maintained and updated,” said Botha, adding that AI agents also cluster related pages, flag outdated content and provide alerts when wiki articles contradict each other.
“AI has the potential to solve the wiki's historical Achilles' heel: manual maintenance overhead,” said Botha. “AI is shifting the wiki from a passive text repository into a self-governing knowledge asset.”
AI builders are also seeing the value of the wiki for their own work. Andrej Karpathy introduced the "LLM Wiki" in the spring of 2026 as a method for building and maintaining a knowledge base with an LLM. The approach works at both the individual and business levels, pulling from a curated list of sources that the LLM can read but not modify. It uses plain markdown files as structured, version-controlled memory for AI agents, in place of vector databases and embedding pipelines. A growing crop of open-source projects (memwiki, Wuphf, Link, mwe-mcp) have landed on the same idea: a folder of interlinked markdown pages, maintained by the agent itself, as the foundation for agent memory.
Building a Wiki People — and AI — Will Use
The secret of a successful wiki is less about getting the software right and more about the rules, guardrails and stewardship that support trust and encourage usage.
“Focus less on the technology and more on governance and adoption,” said Braksiek. “Establish clear ownership, define standards for maintaining content, archive outdated information and make contributing knowledge part of the flow of normal work rather than an extra task.” A wiki is valuable only when people trust that the content is current and reliable, she added.
For a successful wiki, Botha advised taking three steps that prioritize human oversight, governance and clean content:
- Define ownership and review cadences. A wiki is successful only if it is maintained.
- Structure for AI readiness: Design content templates with hierarchies, descriptive headers and explicit tag-based metadata. Clean, modular wiki structures make content is findable and usable by AI models now and in the future.
- Define a clear use case, such as a community of practice or a department where the wiki solves a problem.
Editor's Note: How else are strong knowledge management practices supporting AI?
- How AI and KM Can Power Each Other Forward — Knowledge management has a role in helping define what a future workplace culture should look like in an AI-enhanced world.
- AI-Driven Knowledge Management Turns Repositories Into Intelligent Ecosystems — Knowledge management is undergoing a profound transformation, driven by the rapid integration of artificial intelligence.
- Nicki Usiondek on Why KM and AI Are Better Together — Eli Lilly's Nicki Usiondek shares lessons from an award-winning knowledge management initiative and how it's changed her approach to KM.