Most CEOs are asking their CTO the same question: are we going faster with AI?
It’s the wrong question. It produces reassuring answers, green dashboards, and a false sense of control.
What you’re being told — and what it actually means
Velocity is up. More features are shipping. Cost per feature is down. Your CTO shows you the numbers. Everyone in the room nods.
All of that can be true and still hide an organization that is quietly accumulating invisible debt, producing code no one fully understands, and burning out its best engineers on a validation treadmill they never signed up for.
Output metrics looked great in factories too — right up until the quality crisis hit.
The three questions that actually matter
These aren’t tech metrics. They’re leadership signals.
Are we keeping our commitments?
Predictability is the clearest sign that an organization understands what it’s producing. If your tech team is shipping more but committing less reliably — if deadlines slip more often, if estimates have become fiction — something is wrong underneath the velocity numbers. AI doesn’t fix weak delivery discipline. It amplifies it, in both directions.
Are my engineers getting better — or more dependent?
A high-performing organization that uses AI well becomes more capable over time. Engineers develop sharper judgment about what to delegate, how to verify, where the risks are. If instead your team is becoming faster at producing things they can’t fully vouch for, you’re not building a capability. You’re building a dependency.
Are we building an asset or a risk?
Code that no one can explain six months from now is not an asset. It’s a liability. One of the quieter dangers of AI-generated code is that it can look clean, pass review, and still carry assumptions that nobody made deliberately. The question to ask is not how much your team produced — it’s how much of it they’d be comfortable owning in two years.
The warning sign to watch for
When metrics are green but commitments start slipping. When bugs arrive later in the cycle than they used to. When your team can ship a feature in a day but struggles to describe exactly how it works.
These are signals that AI is being used as a volume accelerator, not as a value lever. The difference matters enormously — and it usually takes six to twelve months to show up in a way that’s impossible to ignore.
The right question
What a CEO should ask their CTO isn’t “are we going faster?”
It’s: “Is the way we’re using AI making us better — or just faster?”
Those are not the same thing. Faster produces more. Better compounds. And in twelve months, you’ll know which one you were actually building.
I work with tech and business leaders navigating exactly this question. If it resonates, I’d be glad to hear how you’re thinking about it. Find me on LinkedIn.