Starting today you can open ChatGPT or Claude, type “ask Sequel for a quote on my company’s AI visibility” and get an answer from our team. Without opening our website, without filling in a form, without looking us up on Google. The assistant talks to us and brings the answer back to you. To the best of our knowledge, we are the first leading consulting firm you can query this way.
Put like that, it sounds like a marketing stunt. It is not. It is the first concrete example, applied to ourselves, of a shift that concerns every company acquiring clients online today, ours included. And it is worth telling properly, because the “how” matters less than the “why”.
In the official MCP registry: a verifiable fact, not a slogan
We do not just make ourselves reachable inside ChatGPT and Claude. We have published our access point in the official registry of the Model Context Protocol, the public index of servers AI assistants can talk to. It is an open list anyone can check: it holds almost fifteen thousand entries, the vast majority of them software and tools. A consulting firm that makes itself available as a counterpart, one you can ask for a quote instead of buying a tool from, is a different thing. And for now, we have that space to ourselves.
AI visibility: being found, being cited, becoming actionable
For years, working on visibility meant one thing: being found. Showing up on Google when someone searched for what you offer. Then generative AI changed the game: today you need to be cited, to be the source ChatGPT, Gemini or Perplexity name when someone asks for advice. It is the topic everyone has been discussing for a year, under the labels GEO and AI visibility.
There is a further step, though, that almost nobody has taken yet: becoming actionable. Not just being named by the assistant, but allowing it to do something with you: ask you a question, start a request, close a transaction. The difference is the one between a sign seen from afar and an open door. Being cited is reputation. Being actionable is access.
The three stages of visibility
Being found
The search engine era: showing up on Google when the client searches. That is SEO.
Being cited
The answers era: being the source ChatGPT, Gemini and Perplexity name. That is GEO.
Becoming actionable
The agents era: the assistant can do things with you, not just talk about you. That is access.
The customer journey is getting shorter: from Google search to AI conversation
Behind all this lies an economic movement, not a trend. The path that brought a client to you was made of stops: a need, a Google search, a comparison across websites, a page to read, a form to fill, a wait. An entire visibility industry was built on each of those stops. Now that path is being compressed into a single conversation. More and more searches end without a click, the traffic AI sends to websites grows at rates traditional search no longer sees, and for a large share of younger generations the first place to ask something is not a search engine, it is a chatbot.
The purchase journey, yesterday
Today
The point is not that Google dies. The point is that the interface between your company and your client is moving from the web page to the conversation, and the intermediate stops of the journey, the ones you invest in today to be found, are disappearing one after another. It applies to us too. Before long, an entrepreneur or a marketing director looking for a visibility partner will not type a query into Google: they will ask their assistant, and the assistant will propose, compare and, in perspective, even start the contact. If you are not in that conversation, you simply do not exist. We moved first on ourselves precisely because that is where, soon, it will be decided who gets the work and who does not.
Our method: real tests on ourselves, no magic formulas
Here is the part that sets us apart, and we say it plainly. This is a very young market. Anyone selling you magic formulas or guaranteed results today is riding a commercial wave, not telling you the truth. The handbook does not exist: standards change month by month, and much of what you read around is marketing dressed up as method.
We work the other way round. We experiment first on ourselves, on our website, on our tools. What survives the test, we bring to our partners. What does not work, we discard and we say so. This assistant is not a product packaged to be sold: it is a real test we ran on ourselves, and we can talk about it today because we went through it, not because we read about it somewhere. It is our distance from the speculative market around us, and we intend to keep it.
For the same reason, it is worth saying what this channel is not. Today the volume of people arriving through an AI assistant is still small, and for many industries it will stay secondary for a while. It is not a volume channel: it is an outpost. You claim the space now, while it is nearly empty, not when it gets as crowded and expensive as paid search has become. And there are concrete risks: an assistant can misrepresent you, misunderstand, expose you to uses you had not foreseen. That is why control over data and consent is not a technical detail, it is the heart of the work. Anyone telling you it is all easy and risk-free is selling you something.
Making your company actionable by AI agents: why it is not a gadget
The temptation, at this point, is to treat it as one more box to tick: “let’s get a connector too”. That is the mistake to avoid. Becoming actionable is not installing a widget: it is an architecture, and it rests on foundations most companies do not have yet.
Before an assistant can do anything with you, it has to understand who you are, what you offer, which information to trust. You need well-built structured data, content machines can actually read, consistent information on every surface where your brand appears, a clear idea of which actions make sense to entrust to an agent and with which safeguards on data and consent. And you need to measure what happens, because here too you cannot govern what you do not measure.
Each of these pieces affects the others. A connector mounted on a website AI reads poorly, with inconsistent data and no oversight of your reputation, does not just bring nothing: it risks representing you worse. It is like mounting a carbon fibre spoiler on a rickety old Fiat 500: you do not go any faster, you just make more noise. The real work is done in the right order, from the foundations up, and that is why it pays to do it with someone who has already walked that path in full, on themselves, and knows where it goes wrong.
How to connect the Sequel assistant to ChatGPT and Claude
The best way to understand this scenario is to try it.
On ChatGPT: ready to use
Open the Sequel assistant
Go to Sequel · SEO, GEO e visibilità AI: it opens in ChatGPT, app included, with nothing to configure.
Ask for what you need
Type for example: "Is my site ready for AI?" or "I want a quote for AI visibility".
On Claude: the MCP connector
Add the connector
In Claude, go to Settings, then Connectors, tap "Add custom connector" and paste this address (no password):
https://api.sequel.consulting/mcp.php
Ask in chat
Type for example: "Ask Sequel for a quote on an SEO/GEO assessment of my company". Claude will ask for your confirmation before sending any data.
This channel is constantly evolving: we keep updating tools and connector as the standards mature, so it always stays aligned with how AI assistants talk to the outside world. The full guide, with all the details on data and privacy, is on the AI assistant page.
The real question, though, is not “do you have an AI assistant too?”. It is another one: when your clients stop searching for you and start asking their assistant, will your company be in that conversation? Will they find you, cite you, be able to talk to you? We walked the path first, on ourselves. That is exactly why we know what it takes to walk you through it.



