#define productivity

“Fast is slow when the wrong road is taken.” — Japanese proverb

There’s a particular kind of sentence I keep hearing lately: “We improved productivity by 40%.” Sometimes it’s 2x. Sometimes 5x. Occasionally, 10x if the LinkedIn post was written after midnight.

And to be fair, something is happening. AI tools genuinely change how quickly many of us can produce things. Code. Slides. Videos. Emails. Documents. Synthetic optimism. The machine is very generous with output.

The problem is that “more output” and “more productivity” are not automatically the same thing. We somehow merged those concepts together somewhere along the way, and now nobody seems entirely sure how to separate them again.

Which becomes mildly inconvenient the moment a leadership team asks:

How do we measure productivity?

Honestly, most organizations don’t really know. But they’re expected to improve it.

Lines of code stopped being useful years ago. The number of meetings was never useful. The number of Jira tickets mostly measures ticket fragmentation. The number of hours online measures… compliance theater, perhaps.

AI now happily generates thousands of words, dozens of wireframes, and entire pull requests before lunch. Quantity has become absurdly cheap. And once output becomes cheap, organizations start confusing activity with value at an industrial scale.

Unfortunately, bad quantity is extremely scalable. The internet already contains enough AI-generated slop to prove this.

Now, I don’t think there’s a clean answer. I also don’t think there’s a universal framework that would be a one-size-fits-all solution. Different teams create value differently. A platform team is not measured the same way as a research team. A startup fighting for survival behaves differently from a mature enterprise trying to reduce operational chaos.

Still, I think there are a few questions worth asking. Once you have answers for those, you can start thinking about metrics.

Is the team solving meaningful problems faster?

Notice the wording carefully. Not:

• Are they producing faster?

• Are they shipping more?

• Are they visibly busy?

A team can dramatically increase output while solving completely irrelevant problems or getting stuck in endless rework cycles. In fact, AI can accelerate this process beautifully. You can now build the wrong thing at unprecedented speed.

Instead, ask: Is the organization becoming better at solving meaningful problems?

• Do customer pains disappear faster?

• Do decisions happen faster?

• Does learning happen faster?

• Do cross-functional bottlenecks decrease?

• Does the organization spend less energy rediscovering the same confusion every quarter?

Those are productivity questions, too. They’re just harder to put into a dashboard. And dashboards, unfortunately, make executives feel wonderfully safe. But these questions will uncover the true drivers of productivity in your organization.

Is cognitive load decreasing or increasing?

This one matters more than many leaders realize, and the more devs I talk to, the more this comes up.

If your engineers now review endless AI-generated code that nobody fully understands, congratulations: you may have automated typing while increasing organizational cognitive debt.

AI lowers the cost of creation. But it does not lower the cost of coordination. Most organizations were never bottlenecked by typing speed in the first place, though. They were bottlenecked by prioritization, decision-making, alignment, trust, and communication. AI often increases the amount of artifacts entering the system while those bottlenecks remain exactly where they were before.

There’s another strange side effect here as well. Talking to AI is comforting. It’s validating, adaptive, endlessly patient, comes up with surprisingly good ideas and gives us the pleasant feeling of competence and momentum. Colleagues tend to do the opposite. They challenge us, misunderstand us, create friction, and expose weak thinking.

Over time, I suspect this may increase silo behavior in organizations. Individual execution speeds up while collective alignment slowly erodes. A company can appear highly productive while slowly losing its internal coherence.

One practical signal to watch for: do people seem mentally clearer or mentally noisier after introducing new tools?

• High-performing teams often feel surprisingly calm internally.

• Low-productivity environments, meanwhile, tend to feel hyperactive. Constant movement. Constant communication. Constant production. Very little actual progress. A bit like a hamster discovering ChatGPT.

What happens to quality over time?

This is where short-term productivity narratives become dangerous. It’s easy to vibe-code a product over the weekend. I’ve done it. It’s genuinely fun. But it’s a completely different story to productionize it properly, extend it significantly six months later, onboard new people into it, or refactor it once reality inevitably changes direction.

Many interventions create a temporary speed boost while quietly degrading maintainability, trust, clarity, or product thinking underneath. Many productivity gains borrow energy from the future.

You can often detect this through second-order signals:

• Are incidents increasing?

• Are teams revisiting rushed decisions more often?

• Is onboarding becoming harder?

• Are senior people spending more time correcting avoidable mistakes?

• Is customer trust improving or eroding?

• Are people becoming more dependent on specific individuals to “untangle things”?

The price of AI-induced tech debt will not magically be cheaper than the traditional kind. We just generate it faster now.

Are you measuring motion or outcome?

This sounds obvious until you sit in enough executive meetings.

Many productivity metrics are actually visibility metrics. They reward observable activity. They create predictable behaviors: fragmented work, over-reporting, excessive status communication, and endless dashboard production.

Meanwhile, some of the most valuable work in organizations remains difficult to see:

• preventing bad decisions

• simplifying architecture

• reducing future complexity

• mentoring

• de-escalating conflict

• creating clarity

• saying “no” early

None of these generates impressive productivity charts. Senior productivity often looks more like subtraction than production.

Good luck turning that into a KPI.

Is the system becoming more adaptive?

This may be the closest thing I personally associate with long-term productivity. And this might be your north-star metric, too.

• Can the organization respond to change without collapsing into chaos?

• Can teams absorb new information quickly?

• Can they rethink assumptions?

• Can they recover from mistakes without weeks of blame games?

To be successful, adaptability matters more than raw throughput. A team shipping 30% slower but making consistently better directional decisions may outperform a “high-productivity” team trapped in perpetual rework.

Especially now, when half the industry seems busy accelerating toward objectives nobody examined particularly carefully. Which, historically speaking, has never caused problems.

So what actually matters?

As you could see, there is rarely a single metric that captures productivity meaningfully. Every organization has its own characteristics, and hence, the metric will differ from company to company. The above questions will help you uncover your own criteria and your own metrics. Look for patterns over time instead of obsessing over a single KPI, and pay attention to the human aspect.

AI amplifies good judgment. But it also amplifies bad judgment in the same way. In a world where production becomes cheap, judgment becomes expensive. Take care of yours.

Food for thought

• When you say your team became “more productive” over the past year, what actually changed underneath the surface: speed, clarity, quality, adaptability… or simply output volume?

• Which activities inside your organization currently look productive but may not create meaningful value?

• Where has AI genuinely reduced friction for your team? And where has it introduced new cognitive load, supervision, complexity or conflicts?

• What are the invisible contributions in your organization that your current metrics don’t capture?

• If your productivity metrics disappeared tomorrow, how would you personally recognize that a team is functioning exceptionally well?