Before You Hire an AI Expert, Recombine the People You Have

Anyone handed the job of driving AI adoption ends up talking about people. Who should own this. Why that person will not move. Before you get there, try one thing.

Pick one AI initiative you are running right now. Picture everyone involved in it, and put a name into each of these four slots.

• Who finds the new ways to use it
• Who turns that into something other people can run
• Who runs it every day
• Who asks "what is this based on?" before it goes out the door

Any slot you cannot fill is probably where the initiative stops. And if the same name goes into all four, it stops on the day that person takes leave.

RELACTIS exists to show you who is standing in those four slots. Twenty-four questions return one of 15 types describing how a person tends to relate to AI. We did not build it to find who is missing. We built it to work out how the people you already have should be arranged. No hiring, no headcount request, just a different arrangement of the same faces.

One person's weak spot is held by the person beside them

Take the first slot, the person who finds new ways to use it. We call that the Pioneer. They are good at turning up a genuinely useful prompt, and it tends to end its life inside their own head or their own working folder. Something used only by the person who found it does not exist as far as the team is concerned.

Trying to fix this through individual effort rarely works. What the result page suggests under "next step" is putting a Systemizer beside them: the second slot, the person who turns it into something others can run.

Systemizers have their own weak spot. Building the thing becomes the point, and they produce a beautifully documented process nobody opens. So before it gets built, name who in the third slot, the person running it daily, is going to use it. That daily operator is weak on the exceptions that fall outside the procedure, which is exactly why they are the one who can tell the Systemizer that a step is awkward in practice.

Look at those three people one at a time and you have listed three weaknesses. Look at them together and each weakness is absorbed by the person next to it. Nobody got fixed, and the places where work used to stop are gone.

Collecting people who know AI does not fill the slots

So this was never a question of expertise. Put three Pioneers together and you get more invention, plus three closed working folders.

The same logic is why we do not start with uniform company-wide training. Give everyone the same session and everyone's existing strength grows a little further in the same direction. But organizations rarely stall because the average is low. They stall because nobody is holding one of the four slots.

The team view is built to raise warnings you cannot get by adding up individual results. Only one person covers a given slot. Nobody is taking results outward. Or the opposite: everyone in the room is too similar. Read the results one page at a time and none of those are visible.

Same people, different purpose, different score

The team view does not stop at one combined number. It splits into four: whether the team fits the purpose, whether it can move the stage you are currently at, how the combination meshes, and how well it holds up against the unexpected. A single total does not tell you what to do next.

Of those, "fits the purpose" cannot be calculated until you supply a purpose. Launching something new scores a team thickest in the people who find things and the people who spread them. Raising quality and reliability puts the heaviest weight on whether anyone is asking what a claim is based on. The identical set of people scores differently the moment the purpose changes.

That is why "hire excellent people" is not an answer. Excellence has no definition outside a purpose.

We don't narrow it to one right answer

Give the org report a purpose and a headcount and it returns three lineups. Three lineups built around the same core would be pointless to compare, so any candidate that reuses the people from an earlier one is deliberately pushed down the ranking.

The three are a range of options rather than a ranking of quality. The product does not pick for you. Comparing them separates the people you genuinely cannot do without from the roles that someone else could just as easily cover, and that separation is the part worth reading. When an organization runs the survey anonymously the lineups stay anonymous too, so it never tells you to pair one named person with another.

What to do with the slot still empty at the end

Even across three lineups, a role can end up in none of them. We call that a white space. It is not something to go hunting for at the start. It is what is left over once the combinations have been worked out.

A white space is still not a hiring conversation. A seven-person department with zero of some type is not evidence that the type is absent from the organization. With a denominator that small, zero happens by chance. Treat a blank as the next place to check.

And filling a blank needs a hire less often than you would expect. A surprising number of slots close simply by saying out loud that a particular person owns that role. People who were never asked are just not doing the thing nobody asked them to do.

The free assessment is 24 questions and takes about four minutes. It starts by showing where you personally stand across those first four slots. Run it across a team and you can see who is in each one. What we want to hand you is not a list of the people you are missing, but a shape you can build with the ones you have.