An agency that is ready for AI search can tell you, on the call, which engines cite your competitors, what they changed on your site in the last quarter to earn citations, and how they measure whether it is working. An agency that isn’t will answer those questions with “we’re keeping an eye on AI” and move on. Most SEO retainers were scoped before answer engines mattered. This is the conversation where you find out whether yours kept up.
The nine questions below are grouped by what they test: visibility, technical foundation, content, measurement, and roadmap. For each one, there is what a strong answer sounds like and the response that should worry you. Related reading: AEO vs. GEO vs. SEO and how to measure your AI citation share.
Why this conversation matters now
Ranking first in a list of links is worth less every quarter that AI Overviews, ChatGPT, and Perplexity answer the question before anyone scrolls. The work that earns a citation in those answers overlaps with classic SEO — crawlable site, real authority, useful content — but adds layers most retainers never covered: structured data kept current, entity grounding, one-question-per-page content, off-site citations, and a measurement loop that tracks whether you are the business being named.
You are not looking for a new category of vendor. You are checking whether your current one moved when the ground did.
Visibility
1. “Which AI engines cite our competitors, and for which prompts?” A good answer comes with examples: the prompts they tested, who got named, and which sources the engines pulled from. It shows they are watching the surface where decisions now get made. Red flag: they have never run your core prompts through ChatGPT or Perplexity, or they treat “AI search” as one thing instead of several engines with different citation behavior.
2. “When ChatGPT describes our business, what does it say — and is it accurate?” Engines carry a summary of your business assembled from across the web. Your agency should know what yours says, whether it is current, and what is feeding any errors in it. Red flag: they have never checked, or they assume it mirrors your homepage. It frequently does not.
Technical foundation
3. “Show me our structured data. When was it last updated?” They should pull up your schema markup, explain which types you are using — Organization, LocalBusiness, Product, FAQPage, Person — and point to a recent change. Red flag: schema was set once at launch and never revisited, or they cannot tell you what is deployed without going to check.
4. “What is our entity presence?” This means a Google Knowledge Panel, consistent name, address, and phone across directories, a Wikidata entry where it fits, and clear links between your site, your profiles, and the people who work there. Engines trust entities they can resolve. Red flag: a blank look, or “that’s not really an SEO thing.”
Content
5. “How many question-targeted pages have we published this quarter, and which questions?” AEO content answers specific questions buyers actually ask, one question per page, with the answer in the first paragraph. A ready agency has a queue of these drawn from your sales calls and support tickets. Red flag: the content plan is still built around keyword volume with no askable question behind each piece, or “content” means a monthly post on a generic topic.
6. “What off-site citations have we earned, and where?” Engines corroborate with third-party sources: industry publications, directories, roundups, forums. Your agency should be earning mentions in the places your engines actually cite. Red flag: link building measured purely by domain rating, with no attention to whether those domains show up in AI answers.
Measurement
7. “How do you measure AI citation share, and what is our trend?” There should be a defined set of prompts, a cadence for checking them, and a record of how often you are the named business over time. Red flag: the only reporting is rankings and traffic. Both still matter, but neither tells you whether an engine recommended you.
8. “How much of our traffic and lead volume is exposed to AI Overviews?” A ready agency has looked at which of your queries now trigger an AI Overview, what that did to click-through, and which pages are affected. Red flag: no awareness that some of your best-performing pages may be losing clicks to an answer box.
Roadmap
9. “What is the six-month roadmap, and how does it split between SEO and AEO?” You want a plan that names the technical fixes, the content queue, the citation targets, and the measurement checkpoints, with rough proportions between them. Red flag: the roadmap is the one they would have written in 2019, or AEO appears only as a line item called “AI optimization” with nothing underneath it.
Score the answers
Give each question 0, 1, or 2 points: 0 for a red-flag answer, 1 for “aware but not acting,” 2 for a concrete answer backed by evidence.
- 14–18: Your agency adapted. Keep going.
- 9–13: Partial. They understand the shift but have not operationalized it. Fixable with pressure and a deadline.
- Under 9: The retainer is behind.
What to do with a low score
Do not fire on the spot. A capable team can close most of this list in a quarter when you make it the priority, and switching costs you months of ramp with the next one.
Send the nine questions in writing and ask for a plan against each. Set a 60-day checkpoint with specific deliverables: schema audited and updated, a citation-share baseline established, the first batch of question pages live, competitor prompt analysis delivered. If those show up, the relationship is worth keeping. If the checkpoint slides or the deliverables come back vague, you have your answer — and you now know exactly what to ask the next agency during the pitch.
The businesses that come through this shift in good shape are rarely the ones that picked the perfect vendor. They are the ones that asked hard questions early and held the answers to a date.
Frequently Asked Questions
How do I know if my SEO agency is doing anything for AI search?
Ask them to run your core buying questions through ChatGPT and Perplexity on the call and show you who gets cited. An agency doing the work will already have this data, know which sources the engines pull from, and be able to name what they have changed on your site to compete. If the answer is 'we're monitoring AI,' with nothing behind it, they aren't doing the work yet.
Should I fire my SEO agency and hire an AEO specialist instead?
Not as a first move. Switching agencies costs three to six months of ramp, and a capable team can close most of the gap in a quarter if you make it the priority. Give them the questions in writing, set a 60-day checkpoint with specific deliverables, and escalate only if the checkpoint slides or the answers stay vague.
What should an AI-search-ready retainer actually include?
Structured data that is audited and kept current, entity work (Knowledge Panel, consistent listings, Wikidata where relevant), a queue of question-targeted pages drawn from real customer questions, off-site citations in sources the engines cite, and a measurement loop that tracks named recommendations over time — not just rankings and sessions.
How much more does AEO work cost on top of SEO?
It is usually a reallocation rather than an add-on. The technical and authority foundations are shared; AEO shifts effort toward structured data upkeep, entity grounding, question-shaped content, and citation monitoring. Any credible scope is priced to your market and goals, so treat a flat rate card as a yellow flag.
How long before AEO work shows results?
Retrieval-based engines like Perplexity and ChatGPT search can begin citing new or improved pages within days of indexing. Google AI Overviews follow your normal crawl-and-trust timeline, typically weeks. Entity-level trust and consistent citation share build over three to six months of steady publishing and cleanup.