8–9 figure operator-founders. AI-curious, eager to evolve.

Companies that are:

• Bullish on systemic change
• Built to move
• Stocked with operators
• Mission-driven — building something that's a net positive for people and the planet

I work hip-to-hip with founders who are bought into the install and have the standing inside their org to make it real.

I want the good guys to win. I’m looking for mission-oriented, kind-hearted professionals.

I don’t take on work if the company’s success would result in a net negative for people or the planet.

First, a conversation to define the problem, the ideal outcome, and the scope together.

Then Phase 1 — Scoping (6 weeks, fixed fee): interview waterfall top-down through the hierarchy. Three deliverables shipped — Business Summary, Roles Org Chart, Ideal Botmap. Automation savings quantified. Gate review at week six.

If approved, Phase 2 — Implementation (milestone-based): 2a POC, 2b department rollout, 2c brain online. Each milestone independently gated. You don't pay for the next until the last one works.

Then Phase 3 — Adoption (retainer): workflow standardisation, data hygiene rituals, AI fluency 1:1s. Success criterion is concrete: your Head of AI ships one new departmental LLM end-to-end without my involvement. When that ships, my work is done.

No — that would provide the wrong return on your investment.

The Decision Engine is a compounding asset. It needs someone inside the company who knows the backend better than anybody else — so they can fix it when it breaks and improve it as the business changes. If I built and operated the whole thing, you'd be paying for a permanent consultant dependency. That's the failure mode I'm explicitly engineered to avoid.

But not all of the build is delegable to a Head of AI. Some of it is, some of it isn't. The split:

Where I work directly with the founder. Certain layers contain strategic IP that shouldn't sit with a Head of AI in their first quarter — or possibly ever. The business summary. The unstated foundational assumptions. The decision-making logic that explains why this company wins. These are founder-altitude documents, often with sensitive commercial detail, and they need the standing of someone who's earned access to them. I work these layers privately with the founder.

Where I work hip-to-hip with someone internal. The roles org chart, the ideal botmap, the workflow standards — these are internal company artifacts your team should have fluency with from day one, not month six. I don't build these for you. I build them with whoever's going to live inside them long-term. They own the backend; I steer the architecture.

Where the Head of AI takes over. Wiring the pipes — permissions, connections, departmental LLMs, dashboard infrastructure. This is the execution layer. The Head of AI runs it against the intent we've already set together.

Different altitudes, deliberately. The model isn't consultant builds, then hands off. It's founder owns the strategic core, internal people own the operational fluency, Head of AI owns the technical execution — and I make sure all three altitudes are pulling in the same direction.

The goal is independence. The day my work is done is the day your team knows the full backend and where every data source lives, without me in the room.

They can — but the gravity of the role pulls them below where the leverage actually is.

A Head of AI is hired to deploy AI on top of the existing stack. Workflow automation, departmental integrations, model selection, infrastructure. Critical work. Real value. But the aperture is narrow by structure — they're executing against a brief, not setting one.

The leverage point sits one level up: redesigning how the company decides, not how the company acts. That's a founder-altitude question. It needs cross-functional context across every department, the standing to ask uncomfortable questions of senior people, and an explicit mandate to challenge the operating model itself — not optimise it.

A peer-level operator coming in from the side gets full leverage. Someone reporting into the operation is structurally constrained from rewriting it.

The model that works: I work hip-to-hip with your Head of AI.

They wire the pipes. I work the founder-altitude layer above — business summary, decision-making process, commander's intent. They execute against the intent. I make sure the intent is right. By the end of Phase 3, they own the system entirely. Different altitude, same direction.

Five risks tracked actively across every engagement, each with a specific mitigation.

1. Data security & containment — the brain touches every department. LLMs cannot bleed into each other.

Mitigation: strict access tiering enforced at the data layer — Notion permissions, Google Drive scopes, API-key isolation per LLM. Finance LLM cannot read HR. Supply chain cannot read marketing customer data. HR cannot read finance. Each departmental LLM has scoped credentials — no shared keys, no shared databases. Every query is logged. Full audit trail of who asked what, when, and what data was touched. Choke points and kill switches at every LLM and at the brain layer.

2. Tooling lag — current models have real limitations.

Mitigation: model-agnostic architecture. Swap providers (Anthropic, OpenAI, local) without rebuilding. As tooling matures, the system inherits the improvements without re-architecture.

3. Adoption failure — most AI rollouts die at the human layer. Data hygiene degrades within weeks if habits aren't changed.

Mitigation: Phase 3 is purpose-built for this. Workflow standardisation, data hygiene rituals, department-lead fluency 1:1s. Success criterion is concrete and measurable: your Head of AI ships one new departmental LLM end-to-end without my involvement.

4. Data hygiene drift — the brain is only as good as the data going in. If project managers stop logging properly, insights degrade silently.

Mitigation: workflow training, logging standards, automated data quality checks built into each departmental dashboard. Weekly reviews per department. "Garbage in, garbage out" made visceral — show teams what bad data does to their own dashboards.

5. Consultant dependency — you don't want to need me forever.

Mitigation: your Head of AI is the handoff target from day one. Phase 3's success criterion is them shipping independently. When that ships, my work is done.

All markdowns, prompts, model configurations, custom software, code, and dashboards are your IP from day one. No vendor lock-in. No consultant lock-in. Everything stays inside your existing stack — Notion, Google Drive, whatever you already use. If I get hit by a bus, your team has full documentation, full access, and a Head of AI trained to run the system independently.