The Intent Map in Practice: A Napkin Before Tax Season
- Jeff Abbott
- Jul 10
- 3 min read
Tom Kessler is the managing partner of a forty-two-person accounting firm in a mid-size Ohio town — the kind of firm where clients are also neighbors, and where the holiday card list and the client list are substantially the same document. Last fall, two of his partners came to the Monday meeting with the same software demo and opposite conclusions. The platform could draft returns, client letters, and advisory memos in a fraction of the time. One partner saw survival: the regional competitors were already using it, and the firm’s margins had been thinning for years. The other saw betrayal: “People bring us their whole lives in a shoebox. Now we’re going to feed that to a machine?”
The argument went three Mondays without moving. What broke it was not a better argument but a diner breakfast, a napkin, and a tool from Chapter 2 of a book Tom’s daughter had given him: the Intent Map. Four quadrants — the book says explicitly it can be drawn on a napkin, and Tom, a literal man, took this as instructions.
Values — what we refuse to compromise
Every client can always reach a person who knows their name. Accuracy before speed, in every season including this one. The shoebox is a trust, not a data set.
Desired Outcomes — the change we’re seeking
Turnaround from three weeks to one during peak season. Partners’ evenings back by March. Shift a third of staff hours from preparation to actual advice — the work clients thank us for.
Guardrails — what we will not do
No client data in consumer AI tools, ever — only the vetted, contracted platform. Nothing leaves the firm without a human who signs their own name to it. And no one loses their job to this in year one — retrain first, then decide what the firm has become.
Metrics — measuring what matters, not what’s easy
Errors caught in review, not just hours saved. Client retention after the season, not just during. How many advisory conversations actually happened. And whether staff would recommend working here, asked anonymously, in April.
The book’s warning sits in how the quadrants connect — values without metrics are just words; metrics without values optimize for the wrong things — and Tom’s partners discovered they’d each been holding half a map. The enthusiastic partner had outcomes and metrics; the resistant partner had values and guardrails. Neither was wrong. Each was incomplete, and each had mistaken the other’s half for the whole of their character.
They adopted the platform in November. But the napkin changed the shape of the adoption: the human-signature rule went into the engagement letters, where clients could read it; the no-consumer-tools rule became a firm policy with teeth; the April staff question went on the calendar next to the tax deadlines. When a client asked in February — as one did — “Is a robot doing my taxes now?”, the answer wasn’t a wince. It was the upper-left quadrant, said out loud: a person who knows your name reads every page that leaves this office. The client, Tom reports, mostly wanted to hear that someone had thought about the question before he asked it.
That is what the Intent Map buys: not the decision, which the pressures of the season would have forced eventually anyway, but the terms of the decision, set down while there was still time to choose them — values made visible before momentum took the choosing away.
Tom is a composite — drawn from conversations with professional-services owners navigating AI adoption, with details changed. The three deadlocked Mondays are real in spirit everywhere.
Try it yourself
One grid, four quadrants, under an hour — Values, Outcomes, Guardrails, Metrics, filled in that order. If you’re deciding with partners, have each person fill their own before comparing. Notice who holds which half of the map; the deadlock is usually two half-maps mistaking each other for opponents.
Comments