The consulting industry loves a methodology. A methodology is a set of steps that, if followed, are supposed to deliver an outcome. The Boston Consulting Group's growth-share matrix is a methodology. The Lean Startup is a methodology. Most AI delivery methodologies are also methodologies: a sequence of steps that look impressive in a slide.

The AI Factory is not a methodology. It is an operating model. The distinction matters.

Why operating models beat methodologies

A methodology can be copied. A team can read the book and follow the steps. The output looks the same on the cover slide but the value at the end of the engagement is rarely there. The reason is that a methodology does not change how the work gets done. It only changes how the work is described.

An operating model changes how the work gets done. The team is shaped differently. The tooling is different. The way deliverables are produced, reviewed, and signed off is different. The economics of the firm running the engagement are different. And, critically, the value that lands on the client's P&L is different.

The shape of an AI Factory operating model

Five stages: Ingest, Analyse, Challenge, Deliver, Embed. Three layers: AI Intelligence, Human Consultancy and Delivery, Institutional Memory. Proprietary frameworks: PAMI, PAVCS, COBIT 5-aligned IT DD. Run from a UK partner team with a wholly-owned India delivery centre at 85% utilisation and a Global Partnership Network behind it.

The five stages are sequential. Ingestion agents read the artefacts at machine speed. Analysis agents map them against the proprietary frameworks. Senior humans Challenge every output. Delivery agents produce the work product. Embed work tracks adoption and value capture against agreed outcomes.

"AI agents do the heavy lifting. Humans hold the quality gate. Value converts to EBIT. Those three sentences are the operating model in one line."

Why the operating model matters more than the model

The foundation models change every six months. Claude, Gemini, OpenAI all keep getting better. The operating model around the model is what stays. The methodology library. The proprietary frameworks. The institutional memory of which prompts work on which problems. The senior judgement that holds the quality gate.

That is why the AI Factory compounds. Every engagement teaches the operating model something. A competitor starting from zero today needs three to five years to catch up on the institutional memory layer, let alone the proprietary frameworks and the senior team experience.

What this means for buyers

  • Ask the vendor what operating model their team runs on. Methodology slides are easy. Operating models are hard to fake.
  • Ask which proprietary frameworks the team brings, and how those frameworks improved between this engagement and the last one.
  • Ask to see a sample of the deliverable produced from a real (or anonymised) data set inside the meeting. AI Factory teams can produce that sample. Methodology teams cannot.

The takeaway. Methodology is a description. Operating model is a way of working. AI changes the way work gets done; the firms that have rebuilt their operating model around AI will compound the advantage. The ones still selling methodology will not.