"Productivity" is coming for your job

AI's greatest impact may be identifying efficiencies in jobs and eliminating useless work. The problem is that most of us do a ton of useless work.

Microsoft's CoPilot is big business. With their built-in- audience, they are probably the best situated to tell us how AI is making us all more productive. The results? Users can read emails faster, write emails faster, and catch up on meetings they missed. Note what they are calling productivity - it's input and messaging. There's no outcome that they're talking about.

This is one of the biggest problems in adopting AI - we have different definitions of success, but we're all pretending that we're using the word productivity the same way. We're not. Marketing is focused on inputs-based productivity (tasks completed) and CFOs are focused on outcomes-based productivity (profits, cost savings, and results). You'll occasionally get lip service to outcomes, but marketing routinely measures what is done, not the impact.

Why does this matter? Executives are making projections and plans based on where they think AI is going - and they tend to run with the benefits of a fully functional agentic or AGI system. Basically - they want their AI to be as good or better than their human workers in all things, not just specific tasks.

But... what if the biggest impact of AI in the short-run is eliminating useless work that we've created over the years to manage employees? What if AI merely pulls the veil aside on how much time employees waste creating "data" for managers to feel like they have something to report?

Salespeople are very familiar with this scenario. If you grew up in a sales culture, you know that you need this many dials to make this many contacts to make this many presentations to make this many deals. Your managers measure dials, meetings, offers - and if your production is low, they point to activity as the cure-all. Technically, this was just shitty sales management, but what else are you going to do? Activity beats non-activity - or at least it appears to. We made salespeople enter nonsense into a database so we could "manage" them if they weren't closing deals.

Apply that to the KPI culture of today. Our managers are running around saying things like, "if you can't measure it, you can't manage it." That's not true - in fact, that quote goes on to say that the impulse to measure and manage is destructive and should be avoided. And yet it continues. We've grown our reporting to such dizzying heights that we're forced to eliminate nuance in order to project certainty. The only data you can be certain about is inputs.

And that's a problem. because the CFO doesn't just care about inputs. They care about inputs to outputs, which means outcomes, which means deals and profitability and cost-cutting. When you tell your CFO you improved productivity by 50%, they don't care until you deliver a 50% cost savings to the bottom line.

That sound dangerous. It reminds me of the VP who stood on stage at a conference and said that referrals were the #1 source of hire, and thus the solution was more referrals. They didn't realize that they were saying the hazy memories of managers were a better source of hire than the entire TA budget. Talk about an own goal! How do you go to your CFO and ask for budget when you just told them that asking hiring managers for referrals is a better choice?

This is why AI is so dangerous in the wrong hands. As we use it more and more, we're finding out it really is easy to do our jobs. And if it is so easy, why are companies paying so much for a human to do it?

As always - my complaint isn't the existence of technology. It's the lies and half-truths we tell ourselves. What's that old saying? Once the tide goes out, you see who isn't wearing a bathing suit.

Productivity isn't about getting into the water.

AI's potential is getting rid of things that humans shouldn't be doing. The misuse of AI is failing to acknowledge that the useless work is how we communicate productivity to the c-suite. Productivity isn't just about doing things faster. It's about doing the right things to deliver value.

Your job depends on getting that definition of success right before you add AI into your company workflow.