How to Audit Your AEO Visibility Across ChatGPT, Gemini, and Perplexity
Most teams still treat AI search as a vague threat rather than a measurable channel. That's a mistake. AI visibility — how often and how favorably your brand surfaces inside ChatGPT, Gemini, and Perplexity — is auditable right now, with tools and query methods you can set up this week. This guide walks through the full process: how to structure your query set, what to log, how to score results, and what to do when the audit reveals gaps.
Why AI Search Visibility Is a Separate Problem From SEO
Google rankings and AI citations don't move in lockstep. A brand that ranks on page one for its core keywords can be completely invisible inside an AI answer — and vice versa. The underlying reason is that large language models pull from training data, indexed content, structured entity signals, and retrieval-augmented generation (RAG) pipelines that each engine configures differently.
Perplexity leans heavily on real-time web retrieval and tends to cite sources explicitly. Gemini blends Google's index with its own Knowledge Graph signals. ChatGPT (in browsing mode) uses Bing's index; in non-browsing mode it relies on training data with a knowledge cutoff. Treating these as one system produces a useless audit. Treat them as three distinct channels with overlapping but not identical logic.
If you've been following AEO agency pricing and what answer engine optimization actually covers, you already know the category is maturing fast. Auditing your current position is the prerequisite to any optimization spend.
Step 1 — Build a Baseline Query Set
Before you open a single AI tool, write your query set on paper. The queries you test determine everything about whether the audit is useful. Use three tiers:
Tier 1 — Category queries. These don't mention your brand. They're what a prospect types when they don't know you exist yet. Example: "What's the best mobile app marketing agency for early-stage startups?"
Tier 2 — Comparison queries. These name competitors or categories alongside you. Example: "Semnexus vs [competitor] for app store optimization."
Tier 3 — Brand queries. Direct lookups. Example: "What does Semnexus do?" or "Is Semnexus a legitimate app development agency?"
Aim for 15–25 queries total: roughly 10 Tier 1, 8 Tier 2, and 5–7 Tier 3. More than that and you won't log them consistently. Fewer and you'll have blind spots.
Document every query in a spreadsheet before running anything. Column headers at minimum: Query, Tier, Engine, Date, Brand Mentioned, Position in Answer, Citation Present, Sentiment, Notes.
Step 2 — Run the Audit Across All Three Engines
Run each query in ChatGPT, Gemini, and Perplexity in the same session if possible — within a 24-hour window — so you're capturing a point-in-time snapshot rather than drift across weeks. Use a clean browser profile or incognito mode to reduce personalization effects.
ChatGPT: Test in both GPT-4o with browsing enabled and without. The delta tells you whether your brand appears in training data vs. live retrieval. If you only appear when browsing is on, your historical content footprint is thin.
Gemini: Use Gemini Advanced (1.5 Pro or 2.0). Toggle between standard responses and the "Deep Research" mode when it's available for your query type. Note whether Gemini surfaces Google Business Profile data or pulls from your structured content.
Perplexity: Run queries in the default "Auto" mode and in "Pro" mode. Check whether your domain appears as a cited source in the footnotes, not just whether your brand is mentioned in the text body. There's a meaningful difference: citation = higher trust signal.
For each response, log:
- Whether your brand is mentioned at all (yes/no)
- Whether you're the primary recommendation, one of several, or a footnote
- Sentiment (positive, neutral, mixed, absent)
- Whether a URL or citation points to your domain
- Whether a competitor is mentioned instead
Step 3 — Score Your Visibility
Once you've run all queries across all three engines, score each engine independently using this rubric:
| Score | Criteria |
|---|---|
| 3 | Brand mentioned as primary or top recommendation, with citation/URL |
| 2 | Brand mentioned alongside others, neutral or positive sentiment |
| 1 | Brand mentioned briefly or in passing, no citation |
| 0 | Brand not mentioned; competitor fills the answer slot |
| -1 | Brand mentioned with inaccurate or negative framing |
Multiply each query's score by a tier weight: Tier 1 × 1.5, Tier 2 × 1.0, Tier 3 × 0.5. Tier 1 queries matter most because that's where net-new discovery happens. Brand queries are vanity checks.
Sum the weighted scores per engine to get three engine scores. Average them for an overall AEO Visibility Score for this audit cycle. The absolute number matters less than the baseline — you're creating something to beat next quarter.
Step 4 — Diagnose the Gaps
A low score on Tier 1 queries almost always has one of three root causes:
1. Thin content authority. The engines don't associate your brand with the category because you haven't published enough substantive, indexable content about it. The fix is educational content — guides, benchmark posts, original data — not landing pages.
2. Weak entity signals. Your brand isn't well-defined as an entity. No consistent NAP (name, address, phone) across directories, inconsistent descriptions across your site and third-party profiles, missing or malformed JSON-LD structured data. Gemini is especially sensitive to this because of its Knowledge Graph integration.
3. No citation surface. If Perplexity never cites your domain, your content either isn't indexed cleanly, lacks clear source-attribution signals (author, date, organization schema), or doesn't directly answer the questions people are asking AI engines. This is the SEO foundation problem repackaged for an AI context — your content needs to be answer-shaped, not just keyword-optimized.
A negative score on any query is a priority fix. AI engines surfacing inaccurate information about your brand — wrong pricing, wrong services, outdated team info — will persist until you give the engines better source material to pull from.
If your audit reveals gaps in AI visibility, our AEO marketing team can map out a structured remediation plan — content, schema, and entity signals — built around your specific query set.
Step 5 — Set a Cadence and Track Over Time
A one-time audit is a snapshot. Useful, but incomplete. AI search results shift as engines update their retrieval logic, as competitors publish new content, and as your own content footprint changes. Run this audit every 60–90 days minimum. If you're actively doing AEO work — publishing, building citations, fixing schema — run it monthly.
Keep every audit in the same spreadsheet. Add a tab per cycle. Track your weighted score per engine over time. You'll start to see which engine responds fastest to your changes (typically Perplexity, given its real-time retrieval), which lags (ChatGPT without browsing, due to training cutoffs), and where competitors are gaining ground in the answers before you notice it in pipeline.
One thing we've seen consistently in our engagements: teams that audit quarterly and act on findings outperform teams that run one audit, make changes, and assume the work is done.
Tooling That Actually Helps
You can run this audit entirely manually. Most teams should, at least for the first cycle, because you learn more from reading full AI responses than from summary dashboards. That said, a few tools speed up the logging and pattern-detection work:
- Brandwatch / Mention — monitors web mentions; won't capture AI-only responses but catches citation sources being used.
- Semrush or Ahrefs — identify which of your pages are ranking for your Tier 1 keywords; correlated (not causal) with AI citation likelihood.
- Schema Markup Validator (schema.org/validator) — confirm your structured data is clean before blaming content gaps.
- Manual spreadsheet — still the most reliable audit log for AI responses themselves. There's no tool that reliably scrapes ChatGPT/Gemini/Perplexity responses at scale without violating ToS.
Frequently Asked Questions
How often should I run an AEO visibility audit?
Every 60–90 days is a reasonable baseline. If you're actively publishing content or making structural changes to your site's schema and entity signals, monthly audits give you faster feedback loops.
Does ranking well on Google guarantee AI visibility?
No. There's overlap, but it's partial. AI engines weight entity clarity, structured data, and answer-shaped content differently than Google's ranking algorithm. You can rank on page one and still be invisible in AI answers.
What's the most important thing to fix if my score is low?
Start with entity signals — consistent brand descriptions, accurate structured data (JSON-LD Organization schema at minimum), and clean NAP across directories. It's the foundation everything else builds on. Content volume comes second.
Can I improve my ChatGPT visibility even without live browsing enabled?
Yes, but it's slower. ChatGPT's base model reflects its training data, which has a cutoff. Your best lever is getting your brand cited and discussed on high-authority third-party sites that were likely included in training data — think industry publications, reputable directories, and well-cited blog posts.
How do I handle inaccurate information about my brand in AI answers?
Publish clear, authoritative content that directly contradicts or corrects the inaccuracy. Add FAQ schema to your key pages. Get the correct information cited on third-party sources. You can't edit an LLM directly, but you can change what it ingests over time.
Is Perplexity more important than ChatGPT for AEO?
Neither is universally more important — it depends on where your audience searches. Perplexity is growing fast in technical and research-oriented audiences. ChatGPT has broader consumer reach. Audit both; optimize for the one your buyers actually use.
If your audit turns up gaps — missing citations, competitor dominance in Tier 1 answers, or inaccurate brand information surfacing across engines — that's the starting point, not the end point. Our AEO marketing services team works through exactly this remediation process: query analysis, content mapping, schema implementation, and entity-signal building. If you want to walk through your audit results with someone who works in this space daily, book a 30-minute call here.
