Hi there!

Welcome to the 30th edition of Work in Beta.

In this edition, we look at a fear a lot of experienced people are carrying right now: if you teach AI to do what you do, do you make yourself replaceable?

So, let’s dive in!

We were working with a finance manager who ran the monthly business review. She’d been doing it for over a decade. Two analysts worked under her: they pulled the numbers from the different systems every month, and she was the one who looked at all of it and said what those numbers actually meant for the business. Is this a real problem, or just a slow quarter? Does it only look scary? The analysts came to her for that call.

Then her company asked us to help them with AI, and we picked the monthly analysis as the place to start. In every session, the analysts came in open and curious. She stayed quiet. She’d answer a direct question and offer nothing more. She was holding back the one thing we needed to make the system work: how the review actually got made.

We understood why. If she handed over the way she works, what would be left for her to do?

Her worry wasn’t foolish. For over a decade, being the person who could make sense of the numbers was the safest job in the building. The review couldn’t go out without her. New analysts joined, learned the systems in a few months, and still ended up at her desk for the final word, because the systems told them what happened but not what it meant. When leadership needs your call before they act, they need you. That isn’t ego. It’s real power. Every rule of thumb she kept in her head was one more reason the team couldn’t close a review without her.

And the AI project came asking for exactly that.

“Building with AI” sounds grand, but it means something simple. An AI system runs on instructions. For it to draft her monthly review, someone has to teach it how that review gets made: where the numbers come from, what to check first, which ratios matter, and what a good summary looks like when it’s done. All of it, written down in plain steps. That written-down way of working is called a playbook.

So our ask, in every session, came down to one line: sit with us and write your playbook. Now her silence made sense. To her, that sounded like writing down her job and handing it over.

Here’s what she couldn’t see yet. Her review was two jobs stitched together, and they’re not the same kind of job.

Job one: making sense of the numbers. Which figures matter, which ratios to run, how a pile of raw data becomes a clean summary. This is method. It can go on paper, step by step, and anyone who follows the steps gets the same summary. An AI follows them in seconds.

Job two: judgement. Reading that summary and telling leadership what it means. This quarter looks fine, but trouble is coming by March, because almost all the growth sits on one big client, and that client has gone quiet. Try writing that down as a step. There’s no rule for “the numbers look fine but something is off.” It only shows up live, with the real situation in front of you.

Her method could go on paper. Her judgement never could.

Put those two jobs next to what we were asking for, and her fear starts to come apart.

The AI needed job one, the playbook, because that’s the only half that can be written down. And that’s the only half an AI can take. The half she was actually scared of losing, the judgement, can’t be handed over at all. There’s nothing to hand. It works only live: in the room on review day, deciding whether to stand firm or ease off when the CEO pushes back.

It’s like football. Give a team the best playbook in the world and they can still lose, because matches aren’t won on paper. Someone has to stand on the sideline, watch the real game, and change the plan while it runs. The AI had her playbook. It still needed her on the sideline.

So yes, the summary part now runs in seconds. What’s left for her is the job she was always best at: telling everyone what the numbers really mean.

What do you do if your role looks like hers? Start small. Pick one piece of work people keep coming to you for again and again. Write down your method for it: where you start, what you check, in what order, what a good result looks like. That’s your playbook. Hand it to your team and your AI. We showed how to write one step by step in Context Document You Should Build when Working with AI.

Then stay the coach. You already do this for your people. Check the AI’s work before anyone acts on it. Make the calls the playbook can’t make. And keep rewriting it, because playbooks go out of date: the business changes, a rule quietly stops being true, and the AI keeps following that rule as if nothing has moved, giving everyone answers that were right last year. AI can run your playbook. It can’t tell when it’s gone stale. You can. That job never ends.

You gave away the steps. You didn’t hand over the game.

AI didn’t come for your job. It came for the half of it that fits on paper.

Write that half down. Stay the coach for the rest. People will stop waiting on you for the summary. They’ll start counting on you for the call.

Most professionals know AI can do more for them. The gap isn’t awareness - it’s knowing where to start, what to change, and how to make it stick.

That’s what we work on through Work in Beta.

For individuals, we run working sessions, not teaching sessions. You bring a real problem from your actual work; we build the solution with you, live. You’re borrowing our learning curve instead of grinding through your own. You walk out with something that works and the muscle to keep going. When you get stuck later, we’re a message away.

You probably have a version of at least one of these:

Bring that. We’ll build it with you.

For organizations, AI adoption is a people problem, not a technology problem. Your teams have the tools, what’s missing is the translation layer between AI capability and daily work. Which processes to redesign, which habits to break, how to build genuine fluency, not just awareness. We help close that gap through hands-on training, process redesign, and deep adoption engagements. Not advisory, forward-deployed.

If any of this resonates, email us at [email protected] / [email protected] and we will figure out how to work together.