AI saves you 11 hours. You keep 4.6.

You were sold 11 hours a week. The receipts say you keep 4.6. Here is where the rest of your time is going.

THE LEVERAGE SIGNAL | Issue #18 | Monday, July 13, 2026

On June 10, Glean's Work AI Institute released the inaugural Work AI Index - a survey of 6,000 full-time digital workers across the US, UK and Australia, co-authored with researchers from Stanford, UC Berkeley, Notre Dame, Emory, UC Santa Barbara, UNC Charlotte and University College London.

The headline numbers look like a win. 87% of digital workers now use AI. 75% say it makes them more productive. They report saving roughly 11 hours a week, over a quarter of the workweek.

Then the number that ruins the story: only 13% say their organization is performing significantly better because of it.

The report found where the hours went. Workers burn 6.4 hours a week on what the authors name botsitting: feeding the model context it should already have, supervising its output, debugging, rerunning prompts that came back thin, and cleaning up confident-but-wrong answers. That is 37% of their AI time - more than the 36% they spend actually producing work with it.

Do the subtraction. 11 hours saved. 6.4 hours of unbudgeted overhead. Net: 4.6.

Nobody is paying you for the 6.4 hours. That is the AI Wage Gap, and it is running through your calendar right now.

The gap was never between the people who use AI and the people who don't. 87% use it. Adoption is settled. The gap is between the people who bank the botsitting and the people who absorb it.

Here is the finding that should reorganize your week. The researchers isolated "high AI achievers" - the workers reporting gains in both productivity and quality. They do almost the opposite of what every AI mandate in your building tells you to do.

They point less AI at their core craft: 38% of their AI time, against 48% for low achievers. They protect the thing they are actually paid for. They botsit more, not less: 40% of their AI time against 33%. They do the verification work deliberately, and they learn the tool's edges from it. And 89% say they know when not to use AI, against 68% of everyone else. Only 33% of all workers are confident they can make that call.

The two paths, sharper than they have ever been.

Path one: you aim AI straight at the center of your job - the analysis, the code, the writing, the judgment you are compensated for - because usage is what gets measured and visible AI fluency is career insurance. Your output volume climbs. Your defensible expertise quietly erodes. 41% of workers now ship work they could not explain if asked. You become the botsitter, and those 6.4 hours stay invisible, unbudgeted and unrewarded.

Path two: you protect the core, aim AI at everything around it, and convert the botsitting into artifacts you own - the context pack, the prompt library with an eval attached, the check that catches the confident-wrong answer before it ships. Same tool. Same 11 hours. Opposite wage trajectory.

The tell that this is a compensation story and not a workflow story: frequent botsitters are 73% more likely to be actively job-hunting. The people holding your AI deployment together are the same people already drafting their exit.

This week, run the subtraction on yourself. One week of honest tracking, one uncomfortable number.

Split every hour you spend with AI into two columns:

If botsitting meets or beats producing (the study's average is 37% against 36%), you do not have an 11-hour dividend. You have 4.6 hours and an unpaid second job.

Then the question that decides which path you are on: can you name one artifact you built this month that kills botsitting permanently? A context pack. A prompt library with an eval. A guardrail that catches the confident-wrong answer before it reaches a client. If you cannot name one, you are paying the tax instead of banking it.

That is the exact thing the BEAST Score measures: whether your AI hours are compounding into leverage or leaking into overhead. 60 seconds, free, no signup. Most people are surprised which side of the line they land on.

You just did the subtraction. Now get the score that explains it.

———————————————————————————

The Leverage Signal is written by Yuri Kruman: 3x CHRO/CLO, AI trainer at Meta, Microsoft and OpenAI, founder of Portfolio Leverage Co. and author of the forthcoming Closing the AI Wage Gap. Reply to this email. I read every one.

📬 Forwarded this? Subscribe free → One signal, one move, every weekday.

]]>