Four ways to answer “why don’t we show up in ChatGPT?”
A client asking this question can be answered four ways: run a free grader, buy a subscription monitoring tool, hand the work to a specialist agency, or deliver a repeatable audit yourself. They are not competing products so much as different jobs. Graders produce a number with no method behind it. Monitoring tracks change over time but does not diagnose. An agency delivers the answer and keeps the relationship. A repeatable audit method is the only one of the four that turns the question into billable work you own.
The short version
| Approach | Answers | Typical cost | Client-ready? |
|---|---|---|---|
| Free AI visibility grader | “Here is a number.” | EUR 0 | No |
| Subscription monitoring platform | “Your position changed this month.” | EUR 29–500 a month | Dashboard, not a deliverable |
| Specialist agency | “Here is the full diagnosis.” | EUR 2,000–5,000 per audit | Yes — under their name |
| A repeatable method you run | “Here is the diagnosis, from us.” | 25–40 unbilled hours to build, or a kit | Yes — under yours |
The four approaches in detail
Option one
A free AI visibility grader
Paste in a domain, get a score. These exist to generate leads for the tool behind them, which is a perfectly reasonable thing for them to do.
When it wins: you want to know in thirty seconds whether a brand has an obvious problem, for your own orientation before a call.
When it does not: the moment a client asks how the number was produced. A grader that will not show you its question set, its engines or its run count has given you a number you cannot defend, and you cannot put an undefendable number in a paid deliverable.
Option two
A subscription monitoring platform
A growing category of tools that track how often a brand is mentioned across AI assistants and chart it over time. Priced per month, usually per brand or per tracked prompt.
When it wins: the client has already done the work and wants to watch whether it held. Monitoring is genuinely the right tool for a retainer where the question is “did this change?”
When it does not: for a one-off diagnosis. Monitoring tells you the score moved; it does not tell you that the engines are routing buyers to a category your client does not describe themselves as being in. It is also a recurring cost against a one-off engagement, and the dashboard belongs to the vendor, not to you.
Option three
Refer it to a specialist agency
Hand the question to a firm that does this full time, and take a referral fee or nothing at all.
When it wins: the engagement is large, technical and outside anything you want to own, or the client needs implementation you are not staffed for.
When it does not: when the client is yours. The agency now has a direct relationship with a company that asked you a question you could not answer, and the revenue that goes with it. On a EUR 2,000 to 5,000 audit, that is the whole engagement.
Option four
Build and run the method yourself
Define a question set, decide how many engines and repeat runs make the result defensible, build the scoring, write the report template, work out what to charge, and then do it again for the next client.
When it wins: always, from audit number two onward. The method is the asset. Once it exists, each audit is about five hours of delivery effort against a fee in the four figures, and the client relationship stays yours.
When it does not: audit number one. Building a defensible method from scratch takes 25 to 40 unbilled hours, and every AI chat session you use to design it gives you a slightly different framework, so the second audit costs almost as much as the first. That gap between the demand existing and the method existing is the entire problem.
Where CITED fits, and where it does not
CITED is option four, pre-built: a folder of seven assets — a question battery builder, a scoring sheet, a client report template, a library of 28 findings with fixes, a commercial pack, a delivery runbook and a licence — that an SEO freelancer or small agency copies once and reuses for every client. It is a delivery system, not software, not a subscription and not a monitoring dashboard. It costs EUR 99 and ships on 15 October 2026.
It is the wrong purchase if:
- You have no clients yet. It is a delivery system, not a way to find people to deliver to.
- You want continuous tracking. Buy a monitoring tool; that is what they are for, and CITED does not replace one.
- You want a guarantee that an engine will cite your client. Nobody can sell that, and the kit’s own scope document says so in writing.
- You would rather not run the audits yourself. Refer the work out instead.
How to judge any of these, including this one
One question separates a measurement from a screenshot: ask how many times each question was run, and what the disagreement rate between runs was.
AI assistants answer the same question differently on different askings. A method that asks once has recorded one point on a curve and cannot tell you whether the result is stable. A method that asks repeatedly and reports its own variance is telling you how much to trust it. If a vendor, a tool or a kit cannot show you that number, it has not measured anything — and that test applies to CITED exactly as it applies to the alternatives.
CITED is operated by Rohde Digital Ventures. cited.rohdeventures.com