AI
How I Judge Team Members Who Use AI: Smart Work or Lazy Paste
Your team has 8 people. Almost all of them use AI now. And you start losing the ability to tell whether the work landing on your desk is your report's actual thinking, or a paste job from ChatGPT.
This is a new dilemma for team leads, and most still judge it by gut feel. That is dangerous. Judge wrong one way and you push out someone genuinely smart, just because they were fast. Judge wrong the other way and it gets worse: you promote someone whose real skill is copy-paste.
I lead four product teams. Plenty of the work that reaches my desk now is done with AI. So I had to build an honest way to evaluate it. Not a fake ban. Because let's be real, banning AI is like banning Google.
Here are the three criteria I use.
Criterion one: ask about context, not output
Producing impressive AI output is easy. Anyone can get AI to write a clean PRD, a slick deck, or a competitor analysis that looks deep. What AI cannot resolve on its own is judgment about context.
The method is simple. I pick one decision from their work and ask why.
Say a PM hands me a roadmap with clear priorities. I do not ask "why does feature A come first". I can answer that myself from the document. I ask things only someone living inside that situation would know. Which stakeholder is going to push back. Which data point still makes them nervous. And if assumption X turns out wrong, does the priority order change.
If they answer fluently and specifically, AI is the tool and they are the driver. If they circle around or fire back with "it's in the draft", that tells me they never really read what they sent.
One context question reveals more than a hundred slides.
Criterion two: check if they verify or swallow it whole
This is the most common failure, and the most expensive. AI will deliver numbers, quotes, and "facts" with total confidence, and some of them are fabricated. People who paste without thinking swallow all of it.
Here is a real example. A team member sent me competitor research with a feature comparison table. I asked them to show me the source for the single most surprising row. From there, two outcomes are possible.
First, they open the source link, and the data is actually there. That means they verified. Second, there is a long pause, then "I forgot to note the link down". That silence is where work habits reveal themselves.
I do not tell them to stop using AI. I tell them to treat AI output like a report from a brand-new intern. Smart, fast, sloppy with details, and sometimes confidently wrong. People who internalize this build a verification habit. People who do not get exposed eventually.
And when they get exposed, the penalty is not "no more AI". The penalty is their reputation taking the hit. That is why this rule has to be agreed on upfront, not weaponized after the fact.
Criterion three: the final call must be accountable to a human
This is the strictest rule on my team. Whatever the tool, the name on the work is a human name.
That means if you submit an analysis for an investment decision and the numbers are wrong because of AI, "the AI said so" is not a defense. You signed it, you own it.
I learned this the hard way, from my own system. An 8-channel content distribution engine that I built and run alone once posted to the wrong account, because the account-matching logic was inverted. The system ran perfectly. It just did the wrong thing very efficiently. There was no AI to blame. I assembled it, and my name was on it.
That is why on my team, AI is perfect for drafting, summarizing, and digging. But final selection, the ship-or-not decision, and accountability in front of others stay with a human. People comfortable with that responsibility look smart using AI. People hiding behind AI look like that too.
The real problem is missing rules, not AI
Team leads are stuck because nobody has agreed on what "good work" means when AI produces 80% of the draft.
I have an example from my own life. Making companies leaner is an old habit of mine, long before AI was trendy: $4M+ per year in savings from various initiatives, from process improvements to more efficient product decisions. Automation was only one of them. That kind of leverage used to require seniority, an engineering team, and expensive systems. Now AI is the sharpest tool for the same habit, and it can be trained into everyone on the team.
But a sharp tool without rules just creates chaos faster. So if you are setting AI rules for your team, start with the three criteria above. Probe the context, require verification, keep accountability with humans. Let your team build out the rest based on their actual work.
If you personally want to build this way of working, we practice together every month in AI Circle, using your real work. And if your company needs help setting AI ground rules and training the team on them, check my corporate page. Pick whichever fits where you are right now.