- January 20, 2026
- Category title
Your AI Strategy Is Missing 80% of What Makes Your Best People Great
Enterprise AI adoption has exploded. By 2025, 78% of large organizations have implemented AI solutions, with spending reaching $37 billion—more than triple the $11.5 billion spent just a year earlier (Menlo VC, Dec 2025). Walk into most of them and you’ll see the same thing: productivity gains happening in pockets. Marketing generates ideas faster. Support handles tickets more efficiently. Engineers code with AI assistance.
All good things. But also surface-level things.
Because while everyone’s focused on making documented processes more accessible, there’s a different problem most organizations haven’t even acknowledged yet. The knowledge that actually matters—the stuff that separates your top performers from everyone else—isn’t sitting in your knowledge base waiting to be fed into an LLM.
It’s in people’s heads (Harvard Business Review, August 2007). Research continues to confirm that 80-90% of organizational knowledge is tacit (Study, Study, Study), yet only 9% of organizations are ready to preserve this knowledge—despite 75% acknowledging it as critical. Meanwhile, 92% fail to consistently capture expertise before employees leave (Deloitte 2020, APQC 2025).
Nabil Keith Durand, MBA, MKM, KM & Organizational Learning Consultant
What We’re Really Talking About
Most enterprise AI today works with explicit knowledge. The procedures, reports, manuals, SOPs—anything that’s been written down and stored somewhere. This is valuable, but it’s also the easy part. It’s already codified. Already accessible, at least in theory.
Tacit knowledge is different. It’s the pattern recognition your best salesperson has developed over hundreds of deals. The intuition your senior engineer has about which technical approaches will create problems six months down the line. The judgment your customer success manager uses to know when a client is about to churn, even when everything looks fine on paper.
This knowledge exists because of experience. It’s contextual, nuanced, and almost impossible to document in traditional formats. You can’t write a procedure manual for “knowing when a prospect is genuinely interested versus just being polite.” But your top salespeople know. And that knowledge is worth something.
In different departments, tacit knowledge shows up in different ways. In sales, it’s understanding why certain prospects convert while others don’t, or knowing which objections are real versus which ones are smokescreens. In customer success, it’s recognizing early warning signs of churn or understanding the real reasons behind feature requests. Engineering teams carry institutional memory about why certain technical decisions were made years ago, what approaches have already failed, and which system quirks never made it into documentation.
Operations might be where tacit knowledge matters most—understanding the informal workflows that make things actually work, knowing who to talk to when processes break down, having lived through enough incidents to recognize patterns before they become problems.
None of this lives in Confluence or SharePoint. It lives in the fifteen years your ops director has been at the company.
Three Levels of Enterprise AI
General GenAI tools like ChatGPT are trained on public data. Useful for general tasks, but they don’t know your products, your customers, or your specific challenges. The output is generic because the training data is generic.
Enterprise GenAI tools like Microsoft Copilot go further. They can access your internal documentation with appropriate security controls, which means the output has organizational context. If you have well-documented processes, these tools can help people find and use that information more efficiently.
Most enterprises stop here. Over 80% of organizations have explored or piloted general AI tools, and nearly 40% report deployment (MIT, 2025). They implement Copilot or a similar tool, see some productivity improvements, and consider their AI strategy handled.
But they’re still only working with that 20% of explicit knowledge. The 80% that’s tacit remains untouched because these tools, by design, can only work with what’s been documented. And here’s what the data shows: these tools primarily enhance individual productivity, not organizational capability (MIT, 2025).
Why This Actually Matters
Organizations facing workforce transitions understand this urgency better than most. Could be retirement waves, could be restructuring, could be rapid growth that requires scaling expertise faster than traditional mentoring allows.
And the timeline is compressed. With 51% of U.S. employees actively searching for or watching for new job opportunities—up from 44% just before the pandemic—and average job tenure down to 3.4 years from 9.2 years in the 1980s, companies face accelerating knowledge loss. The cost? $2.9 trillion annually in voluntary turnover alone, with **41% of organizations rarely or never attempting to collect knowledge from departing employees** (APQC, 2025).
One publicly-listed company I’m aware of had hundreds of experienced employees retiring within a compressed timeframe. Standard enterprise AI could help preserve whatever was already documented. But decades of judgment and context were about to walk out the door, and there was no systematic way to capture it.
They implemented a structured approach to tacit knowledge capture before those employees left. The result was $750K in labor costs saved during restructuring. More importantly, they preserved institutional knowledge that would have otherwise been lost permanently.
The gains showed up in predictable ways—faster onboarding, fewer errors, better customer retention, lower ongoing costs. **Employees typically spend 2.8 hours weekly searching for or requesting information from colleagues, with project delays lasting up to a week in 66% of cases** (APQC, 2021)—inefficiencies that systematically capturing expertise eliminates. But the real value was in preventing the knowledge loss in the first place. You can’t measure what you don’t lose.
Related Posts
Stay informed with the latest updates and insights from the digital world
