Your clients have started asking why ChatGPT never mentions them, and there is no ranking report that answers it. Here is how to answer with evidence, and what the work actually involves.
- The client question is new. The underlying work overlaps heavily with technical SEO you already do.
- The hard part is evidence, because there is no position to report.
- It sells well as a diagnostic first, because the finding creates the scope.
What are clients actually asking for?
Not a ranking. They have noticed that buyers ask assistants for recommendations, that competitors get named, and that they do not. They want to know why, and what it would take to change.
That question has no answer in a rank tracker, which is why it is arriving on agency desks without an obvious deliverable attached.
How much of this is work you already do?
Most of the foundation. Crawl access, structured data, information architecture, content quality and page speed all serve both. The additions are narrower than they look.
| Area | Already in your SEO offer? | What is new |
|---|---|---|
| Crawl access | Yes | Assistant crawlers are separate user agents |
| Structured data | Yes | Agreement with page content matters more |
| Content | Yes | Must be liftable as a standalone answer |
| Identity | Partly | Corroboration across sources is now load bearing |
| Measurement | No | Prompt testing replaces position tracking |
How should this be scoped and priced?
Diagnostic first, then a scoped remediation. A diagnostic is easy to sell because it is bounded and produces the evidence that justifies the larger piece. Selling remediation before the diagnosis means guessing the scope.
Price the diagnostic as a fixed fee and the remediation by the finding count. Avoid promising named placement in any assistant, since nobody controls that and the answers move between sessions.
What can you show a client as proof?
A reproducible score with the failures ranked, a before and after on the specific technical faults, and a record of which assistants name the client for their buying questions over time.
That last one has to be sampled repeatedly rather than measured once, because answers vary. Visibility drift explains why a single check is a sample and not a result.
Common questions
Can I resell this rather than build it?
Yes. We work with agencies on exactly that, including scanning under your own reporting. Talk to us about terms rather than rebuilding the measurement side from scratch.
Is this a separate retainer or part of SEO?
Either works. Most agencies start it as a bounded diagnostic project and fold the ongoing part into an existing retainer once the client sees the findings.
What should I not promise?
Named placement in any assistant, and any specific position. The answers vary between sessions, so commit to the measurable fixes and to tracking, not to an outcome you do not control.
Related reading
Run a free scan and get your reproducible 0 to 100 score with a prioritized fix list. Run the free scan.
Your buyers are already asking AI. This is how you get recommended.
Everything in this article is measurable on a live storefront. The Full Assessment and Roadmap reads your website the way ChatGPT, Gemini and Perplexity read it, and hands you every fix in plain English, ranked by what it is costing you, so the buyers who ask AI get sent to you.