Panos Siozos is CEO of LearnWorlds, a platform powering 12,000+ organizations worldwide. He has a PhD in edtech and 20+ years in e-learning.
We know AI can make people faster. But what we’re discovering is that AI by itself doesn’t make organizations smarter. In some ways, it’s actually making businesses less intelligent.
A thousand experiments do not create capability.
Your business is likely running hundreds or even thousands of small AI experiments right now. People in your team are developing new prompts to varying success; others are trying out agents for their OKRs, and some are automating their inboxes. There’s probably some really good stuff happening there, but how much of it are you aware of? How much of it have you learned something from or done something with?
For most CEOs, the answer is very little. What we’re seeing is that most AI has become a single-player game; those advances in knowledge are getting trapped in one person’s chat history, giving the rest of the business very little to build on.
I believe this is worse than a missed opportunity. An organization that can’t see what its own people know can’t reason from it or build on it, and ends up making the same mistakes in three places at once. At that point, it’s the organization that starts to hallucinate, not the AI.
Businesses need to become more like the Borg.
In Star Trek, the Borg are a collective of individuals that assimilates knowledge from each new member, adding what one has learned to the capabilities of the whole. That’s the part of the Borg businesses should pay attention to. Not in the hive-mind sense, but in the way they assimilate knowledge, distribute it through their people and become relentlessly capable.
AI has given us the ability to pool information at a scale we never could before, but information is not knowledge. Knowledge is what you get when those pieces connect and can add the context that makes them usable. This connected knowledge is what creates capability. Once one person’s discovery is shared, others can build on it so the next person starts from what the last one learned rather than from scratch.
This is how the world of academia works, and it can (and should) apply to companies, too. When you create, capture and connect that knowledge, you build capability both at the individual and the organizational level. That’s more of a modern-day moat than any product feature you’re going to spin up.
But while connecting knowledge may sound like something you lay over the business you already have, it is not.
Do not automate the organization you already have.
The development of knowledge infrastructure requires creative destruction and a return to first principles before AI goes anywhere near it. That may sound scary, but adding AI to a broken organization only reproduces the same problems in a faster, more bloated and more expensive form.
The problem is that the market often encourages you to treat this as a tooling problem. Every new AI capability promises to connect another piece of the business: an AI credit here, a new tier there, another feature layered onto an existing system. But adding more tools does not necessarily create more connected knowledge. You can end up paying to automate individual symptoms while the underlying process remains fragmented. It’s the organizational equivalent of painting over a damp wall: the surface looks better, but the problem underneath is still there.
Save yourself from this by revisiting your processes one by one. Tear them apart and ask what onboarding a client, training a salesperson or launching a product should now mean. And that ‘now’ is important here—it should be different, not the same thing at a different speed.
Rather than assuming a task needs a person, a department or an outside agency, begin with the outcome you need, then decide who or what should own producing it. Sometimes that is a person, though increasingly, it’s a person and a system together.
Then get specific, process by process. Name who owns it, who responds and which system or AI supports it. Once that is clear, AI can help you take the process apart, document it, run it and keep revising it.
None of this works without circulation. Whatever gets learned along the way needs a route back into the shared system. Otherwise, knowledge grows stale at the edges while the center still believes it is current.
Learning has to happen inside the work.
You don’t build this capability by simply sending people to training. Formal training can teach a skill, but it rarely captures the problems people are actually encountering in the work. The learning needs to happen inside the work itself, where new approaches can be tested, refined and shared.
When someone finds a better way to do something with AI, treat it as the start of a loop, not the end of one. It gets captured, tried on a real job rather than a demo and checked against the result it actually produced. Then decide what happens next: discard what doesn’t hold up, refine what does and, when an approach consistently produces a better result, make it available as the new default. The next person starts there rather than from scratch and can improve it further.
Capturing this means taking a problem and its solution and packaging them so someone else can use what was learned without you there, in whatever form fits: a walkthrough, a short course or a simple explainer video. The artifact itself can be disposable. What matters is what it leaves behind: a capability the organization now has and can reuse.
Doing this well is a kind of organizational metacognition: understanding a problem, packaging the answer and getting it to the people who need it in time to use it. And it is built from your own people and your own experts, not bought off a shelf.
That is how single-player becomes multiplayer. One person’s discovery stops living in their chat history and becomes the way the organization works.
To make it worth doing, recognize the person whose prompt became the standard or whose hackathon solution solved a real problem. What gets noticed gets repeated, and what gets repeated becomes part of how the organization works.
Do all this and the organization itself gets more intelligent. Skip it, and you just get faster at being wrong. That is the opposite of intelligence, artificial or not.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
