The AI search landscape moved faster in the last 12 months than the traditional search landscape moved in the prior five years. ChatGPT shipped its shopping experience. Perplexity added price-aware product cards. Claude added inline citations and changed how it weights structured data. Gemini's grounding behavior changed twice. Google AI Overviews changed their citation logic at least three times.
Against that backdrop, an Answer Engine Optimization audit that runs once a year is a snapshot of a landscape that no longer exists by the time you act on it. AEO audits need a quarterly cadence — fast enough to catch the shifts that matter, slow enough that you're auditing a stable state rather than chasing every weekly fluctuation.
Here are the eight checkpoints we run every quarter for clients on our AEO services engagements.
Checkpoint 1: Citation share across the engines that matter for your category
For every client, we maintain a tracked set of the queries that matter most for their category. Each quarter, we run those queries against the AI engines our clients actually need to win — typically some subset of ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews — and we measure citation share.
Citation share is the percentage of relevant tracked queries that cite the brand somewhere in the answer. It's not a vanity metric — it's the closest single number to "are we present in AI answers." We track it relative to the named competitors in the category, not against an abstract benchmark.
The pattern we look for over quarter-on-quarter comparison: which engines are gaining citation share for our client, which are losing it, and which competitors are moving up the citation list. The methodology is documented in detail on the citation share monitoring page.
Checkpoint 2: First-mention position
Being cited in an answer is one thing. Being cited first matters more. AI engines weight the first source they cite higher than the last, both in user attention and in inferred authority signal.
We measure first-mention position separately: of the queries where the brand IS cited, what percentage of those is the brand the first source named. A brand cited 80% of the time but always third or fourth in the citation list is in a weaker position than a brand cited 50% of the time but first when cited.
Improving first-mention position is a content depth and entity-disambiguation problem more than a publishing-volume problem. We cover the work in detail in our annual citation report; the most recent edition is at /reports/state-of-ai-shopping-citations-2026.
Checkpoint 3: Mention rate (vs. citation rate)
Mention rate is the percentage of queries where the brand is mentioned in the answer text — whether or not a clickable citation is attached. AI engines sometimes mention a brand as part of the answer body without linking to the source page. Tracking mention rate alongside citation rate catches the cases where the brand is being discussed but no traffic is being sent.
When mention rate is high but citation rate is low, the on-site fix is typically about schema and source-document structure — making the brand's pages easier for the engine to cite as the source for what it's already saying.
Checkpoint 4: Schema audit and validation
Every quarter we re-validate the structured data on the site against the current state of the relevant schemas. This matters because:
- Schema.org adds and deprecates properties — what was valid 12 months ago may have a more specific replacement now.
- AI engines have shifted which schema types they prioritize. The set Perplexity weights heaviest today is not the set it weighted heaviest a year ago.
- Site changes (template updates, plugin updates, new pages) introduce schema regressions that the original launch never had.
The validation pass uses Google's Rich Results Test, Schema.org's validator, and a manual review of the actual JSON-LD emitted on the page (not the version in the CMS that may or may not be rendering).
Checkpoint 5: AEO content depth audit
The single biggest signal AI engines use to decide who to cite is depth — content that answers a question completely, with specifics, with named entities, and without filler. The depth audit asks, for each priority topic in the client's content portfolio:
- Does the page answer the question the user actually has, or only the keyword phrase?
- Are the named entities (brands, products, locations, services) present and disambiguated?
- Is the answer in a format the engine can lift cleanly — paragraphs that stand alone, headings that match the question form, lists where lists are natural?
- Is there a unique perspective, data point, or example that makes this page worth citing vs. the alternatives?
Thin pages and stale pages are the two biggest losers in AEO. The depth audit produces the per-page priority list for the content team to address before the next quarter's checkpoint.
Checkpoint 6: Source-document availability and accessibility
AI engines can only cite documents they can actually fetch and parse. The accessibility checkpoint covers:
- Are priority pages crawlable by the AI engines (no robots.txt block, no inadvertent noindex)?
- Are AI user-agents allowed where the brand wants citation visibility (ChatGPT-User, PerplexityBot, ClaudeBot, GoogleOther)?
- Is the page render fast enough that the engine's fetch doesn't time out?
- Is the meaningful content in the HTML the engine fetches, or is it only present after a JavaScript render that the engine may or may not execute?
This is the lowest-cost / highest-yield part of an AEO audit. A brand that publishes great content but blocks AI user-agents in robots.txt has solved nothing. We see this exact misconfiguration in roughly a third of pre-engagement audits.
Checkpoint 7: Competitor citation diff
Each quarter we run the same tracked-query set for the named competitors in the category and produce the diff: queries the competitor is cited for and the client is not, and vice versa.
The diff has two outputs:
- Content gap list — topics where competitors are winning citation that our client should be competing on.
- Defensive list — topics where our client is winning that competitors are gaining ground on.
Both go to the content roadmap for the next quarter. This is where the AEO audit hands off into actual editorial planning rather than staying as an analytical exercise.
Checkpoint 8: Brand entity health
The final checkpoint is the broadest: how well does each AI engine understand the brand as an entity?
We measure this by asking each engine direct entity questions — "what does [brand] do," "who founded [brand]," "where is [brand] headquartered," "what does [brand] sell" — and comparing the responses to the actual facts. Hallucinations, omissions, and outdated information are flagged for cleanup.
Entity health is fixable, but only with sustained effort across the entity's source documents: the website's Organization schema, the brand's Wikipedia or Wikidata entry if applicable, the LinkedIn company page, the AP/PR pickups, the founder bios. When all sources agree, the engines converge. When they disagree, the engines hallucinate.
What the audit produces
The quarterly AEO audit produces three artifacts:
- A scorecard — citation share, first-mention position, mention rate, schema validation status, and brand entity health, with quarter-on-quarter deltas.
- A priority list — the top 10 to 20 fixes for the next quarter, ranked by impact and effort.
- A content roadmap update — the topics and angles to commission for the next quarter based on the competitor diff and the depth audit.
That cadence is what makes AEO real work rather than a marketing posture. The AEO audit and SEO audit services pages cover the broader scope; the AEO audit is the AI-specific subset, run quarterly rather than annually because the landscape demands it.
If you'd like to talk through a quarterly AEO audit cadence for your brand, we'd like to hear from you. For the AI-first content side of the work, see AI SEO services.