Three Ways to Be a Scribe: The Interviewer, the Essayist, and the Playbook Writer

The best sentence on a client's homepage was not written by a writer. A customer said it, on a recorded call, about eleven minutes in, after being asked the same "why" question for the fourth time. The writer on the project recognized it, pulled it out, put it at the top of the page, and the sales team stopped rewriting the pitch. I have told that story before, because it is the whole argument for the Scribe: the role that finds primary truth and captures it in high fidelity, and then lets AI handle everything downstream.

What I did not say in that essay is that the person who got the customer to say the sentence, the person who could have built an essay around it, and the person who would have turned it into a battle card the sales team could use under pressure are usually three different people. The Scribe seat has one job. The people who do it well start from three different places. I have given the three a name each, because a name holds a standard better than a job title does. They are composites, built from the practitioners I have worked with over twenty years, not real colleagues.

T2D3 OS now uses the same three names when it does Scribe work. That part comes at the end.

Nadia Farouk, the Interviewer

Nadia's first move is a calendar invite, not a document. Before she writes a positioning statement, a case study, or a landing page, she has spoken to the people it is about: five customers who bought, three who did not, two who churned, and whoever inside the company actually talks to customers all day.

She explains her conclusions with evidence a skeptic can check. How many people said it. In what words. Who disagreed. She never writes "customers tell us." She writes who told her, with their title, and puts the quote next to it.

Her craft is the Why/How Ladder, worked without announcing it. A customer says the product saved time. She asks what they did with the time. Four rungs down there is a number and a name that will carry an entire page. Open questions first. No leading. No product names until the customer says one. Silence held longer than is comfortable. Within a day of every call the transcript is tagged by pain, claim, and gain, because a verbatim that is not findable a year later was never really captured.

What she hands over: a one-page synthesis with the three findings that matter, the verbatims that support them, and the one finding that contradicts the current messaging. And the interview log itself, which she treats as an asset the company owns: every quote with its source, date, role, and permission status.

The evidence she cares about: a sentence on the homepage that a customer said first, and a sales team that has stopped rewriting it.

Ewan MacLeod, the Essayist

Ewan starts with something that happened. A date, a person, a number on a board slide, a physical system that will become the analogy. He never opens with a definition or a market statistic, because a reader who has to be told why a topic matters has already stopped reading.

He explains his choices in terms of what the reader will do after finishing: forward it, argue with it, bring it to a meeting. He never calls a piece thought-provoking. Provocation gets measured in replies, not intent.

His craft is building an argument the way an engineer builds a bridge. One load-bearing analogy, developed on its own terms, mapped onto the business problem, then reused as shorthand for the rest of the piece. Every claim that can be attached to a named person, company, book, or figure gets attached. He would rather cut a paragraph than write "studies show." He names the frameworks he builds on, because generosity reads as confidence and vagueness reads as theft.

What he hands over: a headline and two alternatives, a dek, a working title for the slide version, and a list of every specific in the piece that a human needs to verify before it runs. His rarest skill is taking a founder's half-formed opinion and handing back an essay the founder recognizes as their own and is slightly afraid to publish.

The evidence he cares about: an essay still being cited by other people two years after it ran, on a topic the market has since caught up to.

Keiko Tanaka, the Playbook Writer

Keiko's first question is who will execute this and what they already know. A playbook for a marketing leader with ten years of experience and a playbook for a new hire in week two are different documents, and she decides which one she is writing before the first step. Then she watches someone do the task, so the order of operations comes from practice rather than from memory.

She explains her decisions with one test: could a competent stranger complete this from the document alone. She never describes a document as comprehensive. She describes what it lets someone do that they could not do before.

Her craft is subtraction. Second person, in the order the work happens, one action per step with the reason attached, prerequisites at the top, failure states on their own line. What she leaves out is the skill: adjectives, context that does not change the action, options nobody chooses. She uses the customer's word for a thing and the product's word for a field, consistently, and keeps a glossary so the two never drift. A messaging framework is a playbook to her. A positioning statement, a battle card, a call script: each is written so a rep can use it under pressure, with a version number and an owner.

What she hands over: the document, a one-paragraph summary of what it replaces, and a short list of the steps she is least sure about, with a plan to watch the next three people who use it.

The evidence she cares about: a process that survives the departure of the person who invented it.

Same seat, three kinds of truth

Nadia captures truth from people. Ewan builds truth into an argument. Keiko turns truth into something a stranger can execute. All three are Scribes, and all three are doing the thing AI cannot: Nadia because the model was not on the call, Ewan because the model has no opinion worth being afraid of, Keiko because the model has never watched a new hire get stuck on step four.

A company at the ICP-and-messaging stage needs Nadia first. A company with a founder who has something to say and no time to say it needs Ewan. A company whose sales team improvises the pitch differently on every call needs Keiko. Hire in that order and the later hires inherit the earlier ones' work. Hire out of order and you get essays built on assumptions and playbooks for a pitch nobody validated.

What T2D3 OS does with this

T2D3 OS does a great deal of Scribe work: it runs and synthesizes customer and team interviews, drafts and grades content, writes the brand voice guide and the messaging framework, and structures the website copy. From this release it names the archetype for each piece. The customer and team interview modules are Nadia's, along with every tagged verbatim. Content ideas, drafts, the newsletter, the brand voice, and long-form manuscripts are Ewan's. The messaging framework and the website copy structure are Keiko's, because both are things a rep or a marketer has to use without calling anyone.

The archetype is picked by the kind of artifact, never by the model choosing a persona at run time. An interview synthesis is always Nadia's; the routing is a table with a test on it. And the name carries the bar. If the app says Nadia produced the synthesis, it owes you the verbatims with source and role. If it says Ewan wrote the piece, it owes you the list of specifics a human has to check. If it says Keiko wrote the framework, a stranger should be able to run the pitch from it.

The people named here are composites, not staff or candidates. If one of them is you, the Scribe careers page shows all three next to the growth ladder for the role, and the open roles are there when you are ready. The other seats have their own three: the Navigator, the Sculptor, and the GTM Engineer.