Why I built The Operator's Brief

A while back, a manager, but most importantly a friend, at a local services firm asked me, over a conversation that I had not expected to turn into a memo, if he could implement AI into his firm. "If he could implement" was the wrong question. It's not a question of implementation, but a question of measuring usage once it was implemented to make sure it was the right investment. His team claimed to know AI, but none of them could show him a single artifact that proved it — a memo that had been improved, a draft that had been triaged, a research task that had been cut from three hours to twenty minutes. His read was simple: his bench was telling him what he wanted to hear. The shift was happening to other firms as well. His firm was, in his words, "burning the runway on the wrong fuel."

That conversation is the reason this product exists.

What is actually happening

The Thomson Reuters 2025 Future of Professionals report has 77% of legal, tax, and risk leaders saying AI will have a "transformational" or "high" impact on their work in the next five years. The McKinsey 2025 State of AI report has 78% of companies using AI somewhere and 1% who would call themselves mature. The number that matters is the gap between those two numbers. The use is everywhere. The maturity isn't.

The maturity gap doesn't live at the engineering layer. The engineers know what to do — they have for two years. The gap lives at the manager layer. The senior associate at a mid-tier law firm. The audit manager three years out of partner consideration. The principal broker at a 30-agent shop. The director of operations at a public services company nobody outside the industry has heard of. These are the people whose firms will look very different in twenty-four months, and the question of whether they look different in a good way or a bad way comes down to whether this layer of operators can demonstrate that they are running their work product differently.

The architecture is changing under them. Jack Dorsey is rebuilding Block around what he calls leads instead of managers — directly responsible individuals who own an outcome, build the work, and show their output. He has been clear publicly that the org chart is flattening, and that the middle of the chart is the part that's flattening. He is not the only one. The shape of the AI-native company has fewer layers, more DRIs, more directly-shipped work, and far less status-report coordination.

If you are a senior associate or a mid-tier manager and you are reading this, you already know what I am about to say. The DRI seat is yours to take. The seat goes to the operator who can show, on a Tuesday, that their team's work product is queryable — that a new hire could read the last twenty deliverables in their area and pick up where their predecessor left off, that the partner's edits get inventoried instead of erased, that the prompts that worked are saved and the prompts that didn't are noted. The seat does not go to the operator who used AI more, or who learned more prompts, or who took more training. It goes to the operator whose work is legible.

Why most AI training fails operators

When I started looking at the existing AI training products on offer for the manager layer, I was struck by how universally bad they were. The expensive ones taught prompt engineering. The free ones taught prompt engineering. The Maven cohort taught prompt engineering. The corporate L&D module taught prompt engineering. The vendor demo from the legaltech sales rep — prompt engineering.

The problem with teaching prompts is that prompts are not the lever. The lever is the artifact. The lever is what your team's work product looks like, who can find it, who can learn from it, how it compounds. The prompt is just the keystroke at the moment of asking. A team that has installed real artifact discipline will get useful output from any model, today's and next year's. A team that hasn't, won't, no matter how much they spend on a prompt library.

So the course I built doesn't teach prompts. It teaches operators to make their team's work queryable. Each module installs one habit per week. Document drafting and review with a three-pass rubric. Decision memos that lead with the decision instead of burying it. Delegation briefs that have outcomes, constraints, and decision rights on the same page. Board updates that open with the one number that matters. Budget asks framed three ways so a CFO can argue with the right frame. The Playbook saves the prompts that produced an artifact that landed, so the operator's next month is faster than their last.

That is the durable layer. AI just makes the cost of not having it visible.

What I actually built

The Operator's Brief is the course, plus four other surfaces that earn their slot:

There is Hayes, an AI mentor who is in your corner. He has read your file before you walked in the room. He greets you in your workspace each morning with one observation about where you actually are. He sits in a corner of every page you open and answers questions you'd otherwise have to track me down for. He is built specifically for this product, not a general-purpose chatbot. He knows the curriculum, your progress, and the kind of operator the course is for.

There is the Playbook, which is where the prompts go after the lesson is over. The operator runs a prompt, rates the output, saves a variation, and the Playbook becomes a living artifact of how their work has changed. After three months, the Playbook is the most useful record of an operator's adoption that any firm has — more useful than the certificate on the wall.

There is the Daily Brief, a weekday email and workspace card with one tactical prompt, one micro-lesson, and one open question. Three minutes of input. The operators I most respect treat this the way they treat their morning Bloomberg — not because the content is breaking news, but because the cadence is the point.

There is the Queryable Team Diagnostic, a fifteen-minute survey that returns a memo, not a score. The memo names a specific bottleneck on three dimensions — legibility, artifact density, closed-loop discipline — and points at the one habit to install this week. The memo is forwardable. Champions inside firms use it to start the conversation with their managing partner.

Who this is for, and who it isn't

This is for the director of operations at a 60-person services company whose owner has stopped asking how AI is going and started asking why margins haven't moved. It is for the principal broker who has seen NAR's 88% number — the percent of buyers who transact with the first agent who returns their call — and knows the next twelve months are about response speed. It is for the VP or director of transformation at a Fortune 1000 function whose team is using AI without anyone being able to show what changed. It is for the audit manager at a regional firm who is tired of explaining to the partner why the engagement margin has held flat for three years. It is for the senior associate at an AmLaw 100-200 firm who is two years from partner consideration and wants the firm to know exactly how their book of business has shifted.

This is not for engineers building AI products. It is not for prompt engineers tuning evals. It is not for the AI strategy consultant trying to map a "framework" onto an enterprise. There are good products for all of those people. This one is different — it is built for the operator who has to ship work tomorrow, not for the person whose job is to think about AI.

If you are an operator, the Team plan starts at three seats and $4,200 — $1,400 per seat. Every seat that earns the credential triggers a $100 credit back to the firm, applied at the end of the quarter via Stripe. The incentive aligns: the operators who actually finish the curriculum get rewarded, and the firm pays less for the seats that worked. Operator and Operator+ exist for individuals who want to learn before they bring it to their firm, with plans starting at $599.

What I am asking

Run the fifteen-minute diagnostic. It is free. The memo lands on screen at the end and in your inbox. You do not have to talk to me, you do not have to take a meeting, and there is no sales follow-up unless you reply.

If the memo tracks with how your team actually works, we should talk. If it doesn't, tell me where it went wrong — I read every reply. The product is better because operators tell me what I got wrong, and I would rather hear it from you directly than read it in a survey.

The rest of this is just whether you and I are working on the same problem.

— William Lee, Founder, The Operator's Brief