For those who lead people that have begun to use AI, there is likely an uncomfortable chasm in availability. A lot of activity is clearly revolving around it. But distinguishing between whether that activity is creating real value or just a lot of motion isn't so simple. Because much of what they are producing is technical, it just seems difficult to know from where you sit.
So, the good news is that you can judge it — and without reading even a single line of code to do so. The indicators that discriminate between a team that's using AI well and one that's just busy with it are about behaviour and dialogue, not technology.
If you are going to shove 41 and into the debate; you only talking signal one
This is the fastest tell.
And an AI-using team makes a very good point about what they stopped doing. The drudgery, the boring and circuitous work disappeared in open spaces, leaving them with the more difficult problems that really matter. They also can use something that has improved — not just delays, but extra time for the work only humans can do.
Tools the team talks about when it is busy with it. What is the newest model, which feature just went live, what plugin are they testing out? You are full of passion yet somehow it is difficult to get them to specify what improved as a result.
Outcomes versus tools. Listen to hear for which one gets priority in the conversation.
Signal #2: Are they still able to explain the work?
Ask how something works. You really want an answer if you have a healthy team — somebody takes you through it and knows exactly what was built (even while the AI spits all the words).
In an unhealthy one, it leans toward "the AI did it" with a vague sense that no one is precisely sure how. That's a warning sign. It indicates the team has produced output of which it doesn't completely comprehend, which is acceptable for a disposable experiment and quite hazardous for anything customers rely upon.
What you are checking is ownership. Good teams own the result. Busy teams gave cause to results that own them.
Red flag # 3: How do theyweather the storm
AI makes things feel slower before faster, and there has always been an early period. How a team responds to that spell says everything to me.
A good team is calm about it. They anticipated the slump, they prepared for it and no one is losing their head at how much more difficult things felt for a stretch there. They see it as just a learning opportunity.
A team still in the grip of failing reads that dip as confirmation AI doesn't work, and starts drifting back to the old way. Same experience, opposite interpretation. The interpretation is what validates or invalidates their achievement of the cash out.
Signal four: does learning spread?
Watch how it goes when he gets something useful from only one individual.
In a good team, the trick that works for one guy makes its way to the others within days. There is a common place to leave it, a recurrent custom to hand it around. The unit gets better as a team.
A busy-but-stuck team consists of one-man-experiments by each individual. Five people learn the exact same lessons and make the same mistakes over and over again, because nothing passes from person to person.
The pattern underneath
Then none of these require any technical expertise. A good AI practice speaks of outcomes, owns the work it does, remains steady even with an early dip, and disseminates what it learns. A busy-but-stuck team does the complete opposite on each count.
You're not assessing the technology. You are evaluating your people's way of working, which you already know how to do.
This week: are you hearing more about what improved, or on what shiny new tool?
No comments:
Post a Comment