Working frames · Rachel McBride

Working frames

How I'm thinking about leadership, organisations, and intelligence in the age of AI.

Most AI advice starts with tools. I think it needs to start with position, judgement, and the shape of the organisation instead. These pieces are the frames I'm building in public. They're live: some are settled, some are still sharpening.

Start here

Intelligence at the core, built around people
My updated stance: build intelligence into the centre of how an organisation works, with people in command.
My position on AI
The shortest statement of where I stand: AI can do a lot, but it can't do your position for you.
A firm's position on AI draft
The same stance, one level up. How a firm holds its ground when AI moves to the centre. Personal stance first; this follows.

Bigger frames

My own frames, published while I'm still working them.

The Intelligence-Centred Firm
What changes when intelligence, not hierarchy, becomes the organising centre of the firm.
The same shape, all the way down draft
The same idea at three sizes: a person, a firm, the data itself.
The Verifiable Core
A design principle: the raw record at the core, intelligence on top, and a human owning both ends.
A verifiable core for one
The same principle run on one person: my actual operating structure, the owned core and the disciplines that keep it true.
Where the tool stops
What building Mappa has taught me about where tools help, where they stop, and what still has to stay human.
You're not a job. You're an architecture.
A frame for thinking about a person's practice, judgement, and reputation as one buildable whole.

The Leaders Ledger

What I'm seeing from New Zealand leaders, month by month.

№ 1 · "Clear, and stuck"
After 53 conversations across New Zealand: the four moves that separate the leaders pulling ahead, and why none need a mandate.

Earlier thoughts

Earlier observations and field notes, from when my work centred on workflow redesign. Kept for the record.

The earlier-thoughts archive
Eleven observations and artefacts from early-to-mid 2026. The thinking that led here; it predates my current frames.
As of

Last true: 28 June 2026

v2.1 · "A verifiable core for one" published under Bigger frames: the Verifiable Core principle run at the scale of one person (the seven-part AI setup), the public version of a private working note shared with peers. v2.0 · New lead stance published: "Intelligence at the core, built around people" (now first under Start here; "My position on AI" kept as the why-you-need-a-position companion). Version stamps added to the two stance/ledger pages that lacked them. v1.9 · Plain-English pass: top block now says what this place is for ("How I'm thinking about leadership, organisations, and intelligence in the age of AI" + position/judgement/shape lede); section labels renamed (Start here, Bigger frames); card descriptions rewritten to one plain register. v1.8 · Opening rewritten for the cold reader: subtitle cut to "Stance before tools," lede now leads with the reader's problem (most AI advice adds tools; that's backwards) instead of the three-loops logic; section descriptors de-jargoned (Working Theory, the architecture piece, the firm-position draft). Summary pass: card descriptions and section notes tightened to one plain register ("the same shape" draft, "where the tool stops," the Ledger and archive notes). v1.7 · "Where the tool stops" published — the Verifiable Core in practice, from Mappa's honest limits. v1.6 · The Verifiable Core published (first public statement of the principle; the meaning is public, operator specifics stay gated). v1.5 · Writing consolidated onto ip: earlier-thoughts archive added (11 pieces, framed as earlier thinking); Leaders Ledger № 1 and the Stance piece moved here as current pages (links now local). v1.4 · Opening rewritten as prose (no bullets or dashes); Stance heading set to "everything starts with position"; drafts tightened; em-dashes reduced across the page. v1.3 · three-loops opening. v1.2 · substrate architecture diagnostic live. The IP-spine index; pieces added as they go live.