What Generative AI Actually Does to Productivity - and Why Most Leaders Measure It Wrong
The Productivity Conversation Most Leaders Are Getting Wrong
When CEOs ask me about AI and productivity, they're usually expecting a conversation about time saved. Hours reclaimed. Tasks automated. Headcount rationalized.
Those things are real. But they're not the most important part of what's happening - and leaders who measure only the time dimension are missing the shift that actually matters.
A study published in early 2025 by the Federal Reserve Bank of St. Louis confirmed what I've been seeing in practice:
generative AI is having a measurable,
significant impact on work productivity across industries.
But the nature of that impact is more nuanced than the headline numbers suggest. It isn't just that things get done faster. It's that the nature of the work changes - and that change has implications for how you lead, how you structure your team, and where you personally invest your time.
What Generative AI Actually Does
Generative AI learns from data and produces new content - text, analysis, code, imagery, structured outputs - that didn't exist before the prompt. Unlike earlier automation, which followed fixed rules, generative AI handles open-ended tasks.
It can draft, summarize, synthesize, generate options, and respond to natural language.
That last point matters more than most people realize.
You don't configure it.
You converse with it.
The quality of what comes back depends almost entirely on the quality of what you put in - how specifically you describe the context, the task, and the format you need. Which means the skill that determines your results with AI is essentially a communication skill.
Most leaders already have it. They just haven't applied it here yet.
Where the Productivity Gains Actually Come From
There are three categories where generative AI produces meaningful productivity impact for SMB leaders. They're not equal in importance.
The first is the obvious one: time recovery from high-volume, low-judgment tasks.
Drafting routine communications.
Summarizing lengthy documents.
Generating first drafts of reports, proposals, and presentations.
Building structured data from unstructured inputs. T
hese tasks eat hours every week across every function in your business. AI handles them in minutes. The time that comes back is real and immediate.
Most leaders stop here. They measure AI success by hours saved on these tasks and call it done.
That's leaving the more significant gains untouched.
The second is decision quality.
Generative AI can analyze patterns across more data than any person can hold in working memory, surface options that weren't on the table, generate counterarguments to test assumptions, and compress weeks of background research into a targeted briefing.
The CEO in my network who ran her strategic pivot assumptions through AI before signing a contract - asking it to identify every reason the strategy could fail in six months - made a better decision than she would have made without that pressure test. Not because AI told her what to do.
Because it showed her what she hadn't thought to question.
This is where AI starts to change not just how fast work gets done, but how well.
The third is innovation cycle speed.
Prototyping an idea,
testing a communication approach,
modeling a scenario,
iterating on a product concept
Each of these used to require significant setup time before you could learn anything. Generative AI compresses the distance between hypothesis and feedback.
Your team can run more experiments in a week than they previously could in a month.
The organizations that understand this are not just saving time - they're accelerating the rate at which they learn.
The Measurement Problem
Here's where most leaders make the mistake.
They measure AI productivity by looking at individual task efficiency:
Time to draft an email.
Time to summarize a report.
Minutes saved per meeting.
Those numbers look good in a dashboard and they're easy to track.
What they don't measure: the quality of decisions made with better information:
The ideas that emerged from a faster prototyping cycle.
The strategic risks that got surfaced before they became expensive.
The cognitive load that lifted when routine work stopped consuming the mental bandwidth needed for real thinking.
These are the gains that compound. And they don't show up in a time-tracking spreadsheet.
If you want to measure AI's actual contribution to your business, you need to ask different questions.
"Are your leaders making better-informed decisions faster?"
"Are your teams running more experiments?"
Is the work that was previously crowding out strategic thinking now getting done without that cost?"
Those are the productivity metrics that reflect what generative AI is actually capable of.
What This Means for How You Lead
The productivity shift generative AI creates changes the job description at every level - including yours.
When AI handles the information processing load, the distinctly human contributions become more visible and more valuable.
Judgment.
Relationship.
Context.
The ability to weigh values against outcomes in situations that don't have a clean answer. The ability to read a room and respond to what's actually happening, not just what's on the agenda.
This is the recalibration that's worth leading deliberately, rather than letting it happen by accident. The leaders who navigate it well are the ones who get explicit with their teams about what AI is for, what it isn't for, and how the freed capacity should be redirected. Not as a one-time announcement - as an ongoing conversation about how the work is changing and what that means for how each person contributes.
The Federal Reserve study is useful data. But the more important evidence is what's already happening in organizations that have started this deliberately.
Productivity isn't a number that goes up. It's a shift in what the organization is capable of.
---
If you want to talk through what this shift looks like specifically for your business and your team - that's a good use of an AI Clarity Call.
Source: Alexander Bick, Adam Blandin and David Deming, "The Impact of Generative AI on Work Productivity," Federal Reserve Bank of St. Louis, Feb. 27, 2025.
