AI Strategy · Leadership

Every employee has their own AI. Who is building collective intelligence?

Does your team's next task start from what it learned on the previous one?

By Theodore GeorgedakisThe Growth Igniter18 September 20268 min read

Book a Growth Strategy Meeting See how the Growth Igniter System works

Isolated points of individual learning joining into a shared knowledge network — Many private learning paths. Only one becomes a team capability.

Imagine a team in which two people solve the same problem, with the same AI, in the same week.

One has already found a good method. The other does not know it exists.

The first has done their testing. They have understood what information they need to provide, where the result becomes superficial and what they need to check before using it. That entire small, valuable journey sits within a personal conversation.

A few desks away, their colleague starts from scratch.

Both use AI. Both can become more productive. Yet the company is spending time again on something one of its people has already learned.

Individual progress can coexist with collective repetition.

When we measure AI adoption through licenses, training sessions and deliverables, this loss can easily remain invisible. We see that people are using the tools. We also need to see whether their experience changes how the team operates.

What we measure	What we see	What remains invisible
Licenses	They show access to the tools.	They do not show whether the way the team works is changing.
Training sessions	They show that people have learned the basics.	They do not show which method has become part of the daily process.
Deliverables	They show results and speed.	They do not show whether the knowledge that produced them remains with one person.

One person may be preparing better sales proposals. Someone else may have found how to identify inconsistencies in a report. The essential question is which of these things the next colleague will also be able to do tomorrow, without repeating all the same trials.

That is where a business's collective intelligence is tested.

The private learning path and the cost that goes unseen

AI makes a private learning path particularly easy. The employee explains, corrects and tries again. The result reaches the team as a presentation, analysis or proposal. The knowledge needed to produce it is often left behind.

If this keeps happening, the business may accumulate excellent deliverables and still depend on the same people to reproduce them.

That dependency becomes visible when someone is absent, changes roles or takes on something else. Then the question arises: “How exactly did they do it?”

The cost began much earlier. Every time a colleague solves the same problem again, the company pays again for part of the same learning.

A shared folder of prompts is not enough

A shared folder of prompts is a possible starting point. To become valuable, it needs to convey enough of the reasoning behind the result.

Take the sales proposal. An executive may have discovered that its quality improves when, before drafting begins, they clarify:

• who makes the purchasing decision,
• which problem takes priority,
• what the company can actually deliver.

Where answers are missing, they return to the client with specific questions.

This is a lesson that can change the entire sales process.

For a colleague to use it, they need to understand why these questions come first, how they affect the proposal and when the available information remains insufficient. If they copy only the final instruction to the AI, some of that judgment is lost.

Organizational memory is built when what we have learned influences the next action.

For this to happen, a useful personal discovery needs to go through four steps.

Framework

From personal discovery to organizational memory

Four steps that turn individual learning into a team capability.

Explain it

What problem did it solve? What does it need to work? What criteria do we use to evaluate the result? A short note with a real example and clear limits on its application gives the next person a foundation for using it.

Have a second person test it

A colleague applies the method to a different case, without constant guidance from its creator. The questions they need to ask reveal the knowledge that remained self-evident only to the first person.

Embed it where the work happens

In the sales example, the new questions go into the shared brief and the review criteria into proposal preparation. This way, knowledge appears when it is needed, within the process the team already follows.

Assign someone responsibility for developing it

Someone takes responsibility for gathering failures and revising the method when conditions change. A shared practice also needs a way to be challenged — otherwise, a useful discovery becomes just another old habit.

Knowledge transfer needs space within the work

This journey needs space within the work. Explaining something you have learned, helping a colleague test it and correcting the method takes time.

If the company recognizes only the speed of individual deliverables, knowledge transfer will always compete with “regular work.” It will happen when there is time to spare. And it will stop when pressure increases.

Leadership needs to make this contribution visible:

• who helped others work better,
• which method reduced repeated corrections,
• which discovery improved the quality of an entire team's work.

The benefit also reaches the person sharing the knowledge. Their colleagues test it in cases they themselves had not encountered. They identify weaknesses, add improvements and return a more mature method to them. Knowledge begins to evolve through its use.

You can start at the next weekly meeting

“What did we learn with AI this week that could make someone else's work better?”

Choose one discovery. Have a second person test it. At the next meeting, examine what they were able to use, where they struggled and what needs to change in the shared process.

That can be the start of a learning rhythm that produces something more lasting than a good deliverable.

The next person to take on a task will have a better starting point. A new colleague will find some of the team's experience available to them. The person who made the first discovery will be able to move on to the next problem.

The company becomes smarter when one person's discovery changes everyone else's starting point.

FAQ

What does collective intelligence mean in the use of AI?

It means that one person's discovery changes everyone else's starting point. It is not enough for everyone to use AI; one person's experience needs to become an available foundation for the next, without repeating the same trials.

Why do AI adoption metrics hide the problem?

Because they measure licenses, training sessions and deliverables. They show that people are using the tools, but not whether their experience changes how the team operates. So collective repetition remains invisible.

Is a shared folder of prompts enough?

It is a possible starting point. To become valuable, it needs to convey enough of the reasoning behind the result: what problem was solved, what is needed for it to work and what criteria are used to evaluate it. If only the final instruction is copied, some of the judgment is lost.

What are the four steps for turning a discovery into organizational memory?

Explain it, have a second person test it, embed it where the work happens and assign someone responsibility for developing it, gathering failures and revising the method when conditions change.

What question starts knowledge transfer within the team?

“What did we learn with AI this week that could make someone else's work better?” Choose one discovery, have a second person test it and, at the next meeting, examine what they were able to use and what needs to change in the shared process.

ShareLinkedIn X / Twitter Facebook

Growth Strategy Session