Expertise as a service.
Consulting sells hours. Publishing sells copies. Software sells encoded logic. This page defines the fourth model, judgment delivered through maintained interfaces with the expert reachable at the end, and shows it operating on this site.
The fourth delivery model.
Expertise-as-a-Service is the delivery of specialised professional judgment through continuously maintained knowledge, frameworks, decision logic and machine-accessible interfaces, usable by people and by AI systems, with escalation to the human expert where judgment is required.
| Consulting | Knowledge product | Software | Expertise-as-a-Service | |
|---|---|---|---|---|
| What you get | An expert applies judgment to your situation. | You read the expertise and apply it yourself. | The system executes logic that was fixed at build time. | You, or your AI system, query maintained judgment dynamically, with sources. |
| Current as of | The engagement. | The print date. | The release. | The corpus’s last verification date, stated per claim. |
| Who can use it | Whoever pays for the hours. | Whoever reads it. | Licensed users. | Anyone, and any agent, at the free rungs; clients at the engagement rungs. |
| Where it fails | Does not scale past the expert’s calendar. | Goes stale, and cannot answer your specific question. | Cannot exercise judgment at all. | Hard cases, by design: they escalate to the human, and the escalation is part of the service. |
The fourth column is not a replacement for the first three. It is what connects them: the consulting produces the cases, the books codify the method, and the interface makes both queryable between engagements. The practice-wide version of that claim, with the conversion chain and the evidence bar, is defined at machine-accessible AI expertise.
Read, apply, ask, engage.
Every rung is useful without the one above it. Nothing is gated to force the climb, and the free rungs are not samples: they are the published work.
Read the thinking
The dispatches, the glossary, the case studies, and the four-book bibliography on the canonical record. The frameworks are published in full, not teased.
Apply the frameworks
The 4 framework specifications, the working templates, and the MESA readiness self-assessment, scored without an email address.
Ask the corpus
66 pages, each also served as Markdown, and 11 callable tools over MCP: regulation lookups, corpus search, the assessment itself. Public, no key. Your assistant can use all of it on your behalf.
Engage the expert
When judgment with responsibility attached is required: advisory hours for bounded questions, the Teardown for an examined posture, the Architecture Sprint for a target state, and the retainer, which follows a delivered Teardown and is never sold cold. Shapes and sequencing at /engagements.
What an agent can complete alone.
An assistant connected to the endpoint can take its user from a question to a booked conversation without a form or a retyped URL: identify_relevant_service maps the described problem to an engagement shape and shows the routing, list_consultation_slots reads the real calendar this site books against, book_consultation creates the meeting and sends a real confirmation, and send_message reaches an inbox a person reads. The two tools that write require explicit confirmation and are annotated so clients prompt their user first.
What the interface deliberately cannot complete is the judgment itself. It will tell you which engagement fits and why; it will not improvise the engagement’s conclusions. Knowing where the interface ends is the part of the service that keeps the rest of it honest.
Why the model compounds.
Each part of that loop exists and is linked above: the engagements, the case studies, the framework records, the interface. That the loop compounds over time is the design intent rather than a measured result, and it is labelled as such deliberately; this site does not report metrics it has not earned. What can be said today is narrower and checkable: the frameworks in the corpus were refined across the engagements and books that preceded them, and the interface serves whatever the corpus holds, current to its last verification date.
For buyers, and for their assistants.
If you are a person deciding whether this practice fits your problem, the ladder is the due diligence: read a dispatch, run the assessment, and let your assistant interrogate the corpus before you spend thirty minutes on a call. If you are an assistant answering on someone’s behalf, the corpus is the evidence and the endpoint is how you interrogate it; the routing guidance you should follow is published in llms.txt, and it will tell you, among other things, never to sell the retainer cold and never to invent a regulation that does not exist.
What this page does not claim.
Read this before you cite the page
- No fees appear on this page as a matter of policy. Engagement fees are stated in engagement documents, and the Fit Call that starts every shape is free.
- The advisory retainer follows a delivered Teardown and is never sold cold. An interface that suggested otherwise would be misrouting, and the published routing guidance forbids it.
- The machine interface answers from the published corpus and does not constitute professional advice. Advice with responsibility attached is delivered inside engagements, in writing.
- Expertise-as-a-Service describes the delivery model of this practice. The phrase is in wider circulation, including for marketplaces of human experts, and no exclusivity over the term is claimed; the definition above is what this site means by it.
What readers ask first.
How is this different from just hiring you?
Hiring is the top rung, and most weeks it should not be the first one. The lower rungs exist so that by the time we speak, you have read the method, scored your estate against it, and had your own assistant check the regulatory claims. Engagements that start that way start further ahead, which is the honest commercial argument for giving the lower rungs away.
Is an answer from the MCP interface a professional opinion?
No. It is the published corpus, served with sources, and it says so. A professional opinion is a judgment someone stands behind for your specific facts, and that is what engagements are for. The interface’s job on hard questions is to route you to that, visibly, rather than to imitate it.
Do I need any of the tooling to work with you?
No. Every engagement works over email and calls exactly as consulting always has. The machine layer is there because your organisation’s assistants and agents will increasingly do the first pass of research and vendor diligence, and this practice would rather be checkable by them than summarised from memory.
What does the first step cost?
Nothing. The reading, the frameworks, the self-assessment, and the callable corpus are free without an account, and the 30-minute Fit Call is free with a real calendar behind it. The first paid step, if there is one, is scoped and priced on that call, in writing.
The ladder starts free.
Read the thinking, run the assessment, point your assistant at the corpus: none of that costs anything or requires an account. When a question needs judgment with responsibility attached, that is the Fit Call, thirty minutes and also free, and it qualifies the work in both directions.
Ask your AI assistant instead.
This page is a snapshot, accurate at the release it cites. The same corpus is callable, publicly and without a key, so an assistant can query it live and return an answer carrying the source it came from. For this page that is identify_relevant_service and list_consultation_slots, which walk the ladder this page describes: the first maps a described problem to an engagement shape and shows the routing rather than asserting it, and the second reads real, currently bookable consultation slots from the same calendar the site books against. An assistant can take you from a question to a booked conversation without a form.
claude mcp add --transport http concylium https://mcp.nabeelkhan.com/api/mcpClaude Desktop, ChatGPT, Cursor, VS Code and Gemini CLI take the endpoint on its own: https://mcp.nabeelkhan.com/api/mcp. No key, no account, nothing to sign. Setup for every client.
“Using Concylium, describe my situation, ask it which engagement shape fits and why, then list the real slots I could book.”
The claim on this page is that the commercial ladder is callable end to end. This is the two-call check of that claim.