A while back, sitting with a group of dads at my kid's school, one of them raised a question that had clearly been on his mind for days. How do you upskill your children for a future full of AI?

Telling a kid to master the work first, then move up to oversee AI, is not as simple as it was for us. We were lucky enough to spend a decade doing the real work with no AI at all. We built instincts, judgment, and a feel for the craft by doing it ourselves, over and over. But a child who grows up with AI from day one might skip that step entirely. And if they skip it, what base is left to judge whether the AI got it right or wrong?

That question does not stop at our kids at home. It is the same one every business faces with its own team. A Skill Maturity Rate (SMR) that works in practice shifts the thinking from judging "how good is my team" to observing how the team actually works, as a habit. Not scoring by impression, but looking at real work: what this person can do, and at what level. Can do it alone, can do it with support, or cannot do it yet. Then those results become a development plan per person.

Where AI comes in

Many people fear AI will replace workers. But for team development, AI changes which skills still hold value, not whether people need to be capable at all. Repetitive, rule-based work keeps getting handed to tools. What stays with people is judgment, reading a situation, caring for the customer in front of them, and using AI well as a tool.

So a well-designed SMR should not measure only the old skills. It should add a dimension: how well this person adopts new tools, AI included, to do their own work better. People who adapt become more valuable over time, not just the ones who do the old job faster.

Here is a concrete example. Say you assess the skill "closing a sale on the floor" across four levels. The criteria are not feelings, but things you can actually observe in the moment.

With criteria this concrete, two managers assessing the same person land on roughly the same result, because everyone looks at the same thing, not at a feeling. This is exactly where the AI dimension fits in. Add one more row for "how well they use tools or AI to do their own work," and a single table shows both the existing craft and readiness for what is new.

Here is what that new dimension looks like. Assess the skill using AI to do your own work with the same kind of criteria: observable, not "have you tried it," but whether the work actually got better.

The part to stress is level 3 and up. It does not measure "fluent with AI," but whether they can still tell if the AI's output is right or wrong. Which loops straight back to that business owner's question. Only someone with a solid base can actually oversee AI, because they know what a good result looks like. Skip that base, and you cannot see where the AI went wrong.

How to start

Start by defining clearly what each role must be able to do, then set capability levels in tiers. With that data, managers can see who is ready to grow, who needs support, and which skills to invest in first, especially the ones AI cannot replace.

So what about that father's question? The answer runs on the same principle. Do not rush to let a child hand everything to AI. Let them do the real work first, until they have a base. Then teach them to use AI as a tool on top of it. In the end, the advantage does not go to whoever uses AI best, but to whoever knows what good work looks like, and uses AI to get there faster.

A good assessment is not there to grade people. It is there to show how to move each person forward, whether that person is a teammate or your own child. When work changes this fast, knowing where someone stands and what to add is what separates the ones who scale well from the ones who scale then stall.