The short answer
Lead with one specific, dated finding about the client's own market: the question a customer asked, the businesses the assistants named, and whether the client was among them. Offer a first audit that answers "are we in ChatGPT?" with evidence, then a monthly service that tracks it and shows the work. Promise evidence and a plan, never rankings or revenue.
The meeting you've already had
Clients ask this because they've started using assistants themselves, or someone on their board has. The honest answer today is usually "we don't know", and that is the opening. You don't need to sell the idea that AI answers matter. You need to show the client what the answers say about them.
Open with a finding, not a deck
Before the conversation, run a free check on the client's domain. Pick one question from it that the client's customers would really ask, and write it up in three lines:
- The question, in the customer's words and language.
- Who the assistants named, with the date.
- Whether the client was named, and one thing that might explain the gap (a competitor's comparison page was cited, for example).
A single dated finding is more persuasive than any slide about the rise of AI search, and it can't be waved away as a generic statistic. If the client is named, that's still a finding: now the question is whether it holds, and against whom.
Scope the first audit
Agree three things with the client before you run it:
- Market and language. Where are their customers, and what language do they ask in? A Greek hotel may need Greek and English; a Dutch B2B firm may need Dutch only.
- The decisions that matter. Not every question is worth tracking. Pick the moments where a customer chooses: "which is best for…", "X or Y", "is X worth it", "who can do… near me".
- Who they think the competitors are. You'll compare this with who the assistants actually name. The difference is often the most useful part of the report.
Set expectations you can keep
Say clearly what the audit is and isn't:
| You can promise | Don't promise |
|---|---|
| What the assistants said, on which dates, with the answers attached | A position or rank in ChatGPT |
| Who they recommended instead, and which pages they cited | Traffic or revenue from AI |
| A prioritised list of changes, each tied to the evidence | That a change will cause a recommendation |
| A re-check that shows whether the answers moved | Daily or real-time tracking |
Clients trust an agency that says "here's what we can measure and here's what we can't" more than one that promises a number. It also protects you when an assistant changes its behaviour overnight.
Turn it into a monthly service
AI answers change, so a one-off audit goes stale. The service that keeps its value is monthly:
- a full audit each month on the same questions, so months can be compared;
- a short weekly check in between, with a note only when something moves;
- a plan of changes your team ships, and a re-check that shows what happened afterwards;
- one report the client can read, with the evidence behind every claim.
How you price it is your call. One simple option is to add it as a line inside an existing SEO or content retainer rather than selling it alone, because the fixes are content and site work your team already does.
Objections you'll hear
| Client says | A straight answer |
|---|---|
| "AI answers change every time. How can you measure this?" | That's why we ask the same questions repeatedly and only report movement that holds across answers and assistants. |
| "Can't I just ask ChatGPT myself?" | You can ask once, in one language, and get one answer that may change tomorrow. The audit asks the questions your customers ask, repeatedly, and keeps the evidence. |
| "Is this just SEO?" | Partly. Google says its AI features rely on the same foundations as Search. But ChatGPT, Gemini and Claude use their own search and their own judgement, so we measure them directly. |
| "What if we're not in there at all?" | Then the first report shows who is, which pages the assistants lean on, and where to start. |
Doing this with Scope
Scope's free check gives you the opening finding for any client domain. Studio and Agency plans run the monthly audits and weekly pulses per client, with white-label reports, a permanent link for each client and a Portfolio view across your roster. See plans and pricing.
What this can't tell you
- The scene above is illustrative. It does not describe a real client or agency.
- An audit is a dated sample of AI answers; it cannot guarantee what an assistant will say next.
Sources
Primary sources first. Research papers are cited with their limits. Every source was re-opened on the date shown.
- 1Scope methodology
The InnerView · checked 25 September 2026
How Scope defines recommendation rate and visibility, compares audits, repeats questions and states its limits.
- 2Scope product facts and limitations
The InnerView · 24 September 2026 · checked 25 September 2026
Supported assistants, free check scope, monitoring rhythm and stated limitations, including that Google AI Overviews and Perplexity are not currently measured.
- 3Optimizing for generative AI features in Google Search
Google Search Central · 10 July 2026 · checked 25 September 2026
For Google, AEO/GEO work is SEO. Rejects llms.txt, content chunking, AI-only rewrites, special schema and inauthentic mentions as requirements.
Get the next guide by email
Short emails on how ChatGPT, Gemini and Claude recommend businesses: the new piece, one real answer we saw, and one fix you can make. Unsubscribe any time.