Doc JC-01AI process automationUS · Canada · UK · TaiwanCBAP® · PMP® · n8n Verified CreatorRev 2026.08

An internal AI system for a small consulting firm, built on the methods you already wrote down

A consulting firm sells its method. In most small firms, the method is already written down, in years of methodology and project files. The problem is getting at it. A consultant with a client question still digs through a shared drive, or pastes the background into a chatbot and explains it all again.

Some people call what fixes this an internal AI operating system. What I build is narrower and easier to check: an assistant your consultants ask in plain language, answering from your firm's own documents.

Twenty years of method, made askable

One client is a brand consultancy in Taiwan with two decades of methodology fully written down. They still answered every client question by hand. In their words, the intelligence is all there, it just isn't connected.

The assistant I am building lets their consultants ask in plain language. It answers from the firm's own methodology and project files, and shows which document each answer came from. The work is still in build, and the record of work says so.

Every answer shows where it came from

A consultant has to be able to stand behind what they tell a client. That is why the assistant shows which document each answer came from, so the consultant can open it and check. In the document systems I build, a question the files cannot answer gets "no record" instead of a guess, and I test for that on purpose.

The judgement stays with your consultants

On day one, this client set one rule: human judgement stays central. That shaped the whole build. The assistant finds and drafts. Deciding what to recommend to a client is still the consultant's job, and the system is designed around that.

You write the test it has to pass

Before you rely on it, you give me real questions your team asks, and you decide what counts as a correct answer. How that test works, and what the system can and cannot do across a large archive, is on the document archive page.

A one-week start inside the AI you already pay for

If your firm's documents fit inside a tool you already use, such as ChatGPT, Claude, Gemini, or Copilot, the Document Starter is $500 fixed and takes one week. I sort the documents, load them, test the result on thirty questions your team actually asks, and give you a written report on every answer.

A system built for your firm

If the archive goes back years, or you need search built for the firm instead of living inside a subscription, it is a separate project with its own quote and a monthly fee to keep it running. Both sizes are on the services page.