The 7-Step AEO Audit Every Marketing Team Should Run

First published: 22 June 2026 · Last updated: 22 June 2026

1

Brand visibility check

Are AI engines finding your brand at all? Run 10 branded and category prompts on ChatGPT, Perplexity, AI Overviews. Log who gets cited.

2

AI crawler access check

Is GPTBot, ClaudeBot, PerplexityBot, Googlebot allowed in robots.txt? Most SG sites silently block one or more.

3

Schema baseline

Run Rich Results Test on 5 key pages. Confirm Organization, FAQPage, Article schema present and validated.

4

Content answerability

Are pages structured as scannable answers? H2s as questions, 30-50 word answer leads, short paragraphs, clear hierarchy.

5

FAQ coverage

Do top 10 commercial pages have at least 4 FAQ items each, marked up with FAQPage schema? Most miss this.

6

Citation comparison

Who is currently being cited for your top 20 commercial queries? What format do their cited pages share?

7

Tracking setup

Quarterly manual citation log + GA4 referrer tracking for chat.openai.com, perplexity.ai, gemini.google.com.

If you are a marketing manager at an SG SME, you have probably been told you "need to do AEO" by someone in the last six months. You probably have not been told what an AEO audit actually involves or how to run one without hiring an agency at SGD 5,000+. This article fixes that. It is a 7-step audit you can run yourself in roughly half a day, with no code knowledge required, that produces a prioritised fix list ready to hand to your dev team or your agency. This is the decision-maker version of the audit. We have a deeper 12-question practitioner version on BestSEO covering technical implementation depth (AI SEO audit workflow) for your dev or SEO specialist to read after this. The framework here is the structured walkthrough every marketing leader should run quarterly so you stay on top of your AI visibility without depending on an agency to tell you whether things are working. Our piece on AEO vs SEO in Singapore covers the foundational difference between the two disciplines.

Step 1: Brand Visibility Check (45 minutes)

The first question you need to answer is the simplest: do AI engines know you exist? You would be surprised how many SG SMEs are completely invisible to ChatGPT and Perplexity for queries they should own. Open three tabs: ChatGPT (chat.openai.com), Perplexity (perplexity.ai), and Google with AI Overviews enabled. Run these 10 queries, log who gets cited:

1. "[Your brand name]" (does it appear at all?)
2. "What does [your brand] do?"
3. "Best [your category] in Singapore" (e.g. best digital marketing agency in Singapore)
4. "How much does [your service] cost in Singapore?"
5. "[Your category] for [specific buyer], e.g. SEO for healthcare clinics in Singapore"
6. "[Your top 5 competitor 1] vs [Your top competitor 2]" (are you in the comparison or invisible?)
7. "Top 10 [your category] companies in Singapore"
8. "How to choose a [your category] in Singapore"
9. "[Your category] reviews Singapore"
10. "[Your most-bought product or service] price Singapore"

Use a simple spreadsheet. Columns: query, ChatGPT cites you Y/N, Perplexity cites you Y/N, AI Overviews cites you Y/N, who got cited if not you. The pattern that emerges is your visibility baseline. If you are cited zero out of 30 (10 queries x 3 engines), you have a foundational gap. If you are cited 5 to 10 out of 30, you have a typical SG SME baseline. Above 15, you are doing well.

Step 2: AI Crawler Access Check (15 minutes)

The most embarrassing AEO mistake is a site that has accidentally blocked the AI engines from crawling it. This happens more often than you would think; old robots.txt rules, dev team caution, security plugins. The audit step is simple. Open `[yourdomain.com]/robots.txt` in your browser. Look for any lines that mention these user agents:

• `GPTBot` (OpenAI's crawler for ChatGPT and ChatGPT Search)
• `ClaudeBot` and `Claude-Web` (Anthropic's crawler for Claude)
• `PerplexityBot` (Perplexity's crawler)
• `Google-Extended` (Google's crawler for AI Overviews and Bard/Gemini training, separate from Googlebot)
• `Applebot-Extended` (Apple Intelligence crawler)
• `Bingbot` (Bing's classical crawler, also feeds Copilot)

For each, you want either no mention (default allow) or an explicit `Allow: /` rule. A `Disallow: /` rule for any of these means you are invisible to that engine. Fix is a one-line robots.txt change. If your robots.txt is empty or missing, that is fine; default behaviour is allow-all. The danger is the explicit blocks. Some agencies and security plugins add these "for safety" without realising the consequence is invisibility. Get this fixed first because every other AEO investment is wasted if the engines cannot read your site.

Step 3: Schema Baseline (45 minutes)

Schema markup is the structured data that helps AI engines extract and cite your content with confidence. We covered the strategic case in GEO schema markup for AI citations; this is the audit version. Pick 5 representative pages: your homepage, your top service or product page, your top blog post, your top FAQ-heavy page, and one location or service-area page. For each, do this:

1. Open Google's Rich Results Test (search "Rich Results Test").
2. Paste the page URL.
3. Look at "Detected items". Note which schema types are present (Organization, LocalBusiness, FAQPage, Article, Product, Service).
4. Note any errors or warnings.

Build a simple table: | Page | Organization | LocalBusiness | FAQPage | Article | Product/Service | Errors | |------|--------------|---------------|---------|---------|------|--------| Fill it in for the 5 pages. The gaps are obvious. If your homepage has no Organization schema, that is your priority 1 fix. If your service pages have no Service schema, that is priority 2. If your blog posts have no Article or FAQPage schema, that is priority 3. Hand this table to your dev team with a brief: "ship the missing schema for these 5 pages first, then roll the same templates across the rest of the site."

Step 4: Content Answerability Check (45 minutes)

AI engines preferentially cite content that is structured as scannable, extractable answers. Walls of marketing prose are bad citation candidates. The audit question is simple: are your top pages structured for answer extraction? Pick the same 5 pages from Step 3. For each, score yes or no on these criteria:

• Question-form H2s. At least 3 of the H2 headings on the page are phrased as questions ("How does X work?", "What is the cost of Y?")
• 30 to 50 word answer leads. Each major H2 is followed by a 30 to 50 word direct answer in the first paragraph (rather than a 200-word warm-up).
• Short paragraphs. No paragraph longer than 4 sentences. AI engines bail out on dense walls of text.
• Clear hierarchy. Logical H1 to H2 to H3 structure, not a soup of equally-weighted headings.
• TLDR or summary. A short summary or quick answer block near the top of the page.

A page scoring 4 or 5 is well-structured for AEO. A page scoring 2 or below needs restructuring. The fix is content work, not dev work, and is usually achievable inside a single afternoon per page.

Criterion	Why it matters	Pass threshold
Question-form H2s	AI engines extract answers to questions, not topics	3+ per top page
30-50 word answer leads	Featured snippet and citation pull length	Every major H2
Short paragraphs	Dense text drops citation probability	Max 4 sentences each
Clear hierarchy	Heading structure signals content map	H1, H2, H3 logical
TLDR or summary	Quick-extract block for AI summarisation	Within first 200 words

Step 5: FAQ Coverage Check (30 minutes)

FAQPage schema is currently the highest-leverage AEO move. AI engines actively pull from FAQ markup because it is pre-structured Q&A format that maps cleanly to what they need to generate. Yet most SG SME sites have FAQ blocks on only their homepage, if at all.

Audit your top 10 commercial pages (highest-traffic service pages, top product pages). For each:

1. Does the page have a visible FAQ section with at least 4 questions and answers?
2. Is the FAQ marked up with FAQPage schema (check via Rich Results Test)?
3. Are the questions mined from real customer queries (sales call FAQs, support tickets, People Also Ask) or made-up filler?

A page that fails any of these three is a clear fix. The FAQ block is one of the cheapest content additions you can make. Mining the questions takes 1 to 2 hours per page (review last 30 days of sales calls, last 50 support tickets, top PAA results). Writing the answers takes another 1 to 2 hours. Adding the schema takes 0 hours if your site is templated; otherwise 30 minutes per page for a dev. Our piece on People Also Ask optimisation covers the question-mining process in depth.

Step 6: Citation Comparison (60 minutes)

Now look at who is winning what you should be winning. For your top 20 commercial queries (the ones from Step 1, plus 10 more high-intent ones), log which pages and brands are getting cited in ChatGPT, Perplexity and AI Overviews. Look for patterns.

What format are the cited pages in? Long-form guides? Comparison articles? Listicles? Service pages with FAQs? Product detail pages? The format pattern is your prescription for what to build.

What domains are dominating? If three or four domains are consistently cited across your top queries, those are your AEO competitors. They have done something right. Visit each one and look at:

• Do they have visible FAQ sections?
• Do they have explicit Q&A content?
• Are their pages structured with question-form H2s?
• Do they have author bylines and bio pages?
• How long is their content?

You will see a pattern. Replicate it. This is not about copying; it is about matching the format conventions that the AI engines have learned to trust in your category. If every cited page in your category is 2,500 words with FAQ schema and a clear author byline, your 600-word service page with no FAQ is structurally disqualified from citation.

Step 7: Tracking Setup (15 minutes)

You cannot improve what you do not measure. AEO tracking is still primitive in 2026 (no native rank-tracker for AI engines), but two simple setups give you a baseline.

Manual quarterly citation log. Set a calendar reminder for the first week of every quarter. Re-run your top 20 queries on ChatGPT, Perplexity and AI Overviews. Log who gets cited. Compare to last quarter. This is the only true measure of your AEO trajectory.

GA4 referrer tracking. AI engines that send actual click-throughs (Perplexity does, ChatGPT Search does, AI Overviews sometimes does) will appear in your GA4 referral reports. Filter or annotate for these specific referrers:

• chat.openai.com
• chatgpt.com
• perplexity.ai
• gemini.google.com
• copilot.microsoft.com

Track the trend. Even small positive trends here are early signals that your AEO investment is working. We covered the deeper measurement question in measuring GEO performance with the right tools.

What to Do With Your Audit Results

After the 7 steps you should have a prioritised fix list. The typical SG SME site comes out with 5 to 15 specific actions. Group them three ways:

Quick wins (under 4 hours of work each):

• Fix robots.txt to allow AI crawlers
• Add FAQ blocks to top 5 commercial pages
• Restructure 3 service pages with question-form H2s
• Add 30 to 50 word answer leads to existing H2s

Medium projects (1 to 5 days of work):

• Ship Organization, LocalBusiness, FAQPage, Article schema sitewide
• Build out FAQ content on remaining commercial pages
• Add author bio pages and proper Article schema bylines
• Restructure top 10 pages for answer extraction

Strategic investments (ongoing):

• Quarterly AEO audit cadence
• New blog content built around question-form titles and answer-first structure
• Manual quarterly citation tracking
• Content team training on AEO writing patterns

Do the quick wins this month. Schedule the medium projects across the next quarter. Build the strategic investments into your standing scope.

1

This week: unblock crawlers

If any AI bot is disallowed in robots.txt, fix today. Every day of invisibility compounds.

2

This month: quick content wins

FAQ blocks on top 5 pages, question-form H2s, 30-50 word answer leads. Content team work, no dev needed.

3

This quarter: schema sweep

Organization, LocalBusiness, FAQPage, Article schema sitewide. 8-20 hour dev project, big visibility payoff.

4

Ongoing: cadence + tracking

Quarterly re-audit, monthly GA4 AI-referrer review, AEO standards in every new content brief.

Frequently Asked Questions

How is an AEO audit different from a regular SEO audit?

A regular SEO audit checks ranking factors for classical Google search: backlinks, page speed, on-page keyword usage, technical crawlability. An AEO audit checks visibility factors for AI answer engines: structured data depth, content answerability, AI bot crawler access, citation patterns. There is overlap (technical crawlability matters for both) but the priorities are different. AEO weights schema, FAQ coverage and answer-format structure more heavily than backlinks or keyword density.

Do I need an SEO agency to run this audit?

No. The 7-step decision-maker version is designed for an in-house marketing manager to run with no code knowledge. You will need a dev team or agency to implement some of the fixes (especially the schema work), but the audit itself is self-service. If you would rather hand the whole thing over, that is what our AI SEO service is for. For a deeper practitioner-level audit with technical implementation depth, our piece on the AI SEO audit workflow at BestSEO is the next-level read.

How often should I re-run the AEO audit?

Quarterly is the right cadence. AI engines update their citation logic regularly, your competitors move, and your own site changes. A quarterly audit catches drift before it costs you visibility. For high-stakes commercial sites or sites in fast-moving categories, a monthly mini-audit (Steps 1, 2 and 6 only) is also worth doing.

What if my brand is invisible across all 30 audit queries in Step 1?

That is more common than you would think for SG SMEs. The good news is the fix path is clear: ship Tier 1 schema (Organization, LocalBusiness), add FAQ blocks to your top 5 commercial pages, ensure robots.txt allows AI crawlers, and start producing 1 to 2 question-targeted blog posts per week. Most brands see their first AI citations within 60 to 90 days of starting this work consistently.

Does AEO work cannibalise my classical SEO?

No. AEO improvements (better schema, clearer answer structure, FAQ depth, answer-form H2s) almost always also improve classical SEO performance because Google's traditional ranking algorithm rewards the same signals. The two disciplines are largely additive. The only exception is if you produce content that is purely AEO-optimised filler with no genuine value; that hurts both. Quality content with AEO-friendly structure is the goal.

Should I worry about AI engines citing my content without driving traffic?

Worth thinking about, not worth panicking over. AI engines do increase zero-click impressions, but they also drive direct traffic (Perplexity in particular sends meaningful click-through), and brand mentions in AI answers compound brand recall even without clicks. Our piece on zero-click search strategy covers how to win when no one clicks. The brands that refuse to optimise for AI engines because "they will not click anyway" are choosing invisibility, which is worse than zero clicks.

Related reading

• AEO vs SEO in Singapore: how the rules differ
• People Also Ask optimisation for SG businesses
• AI SEO audit workflow (BestSEO practitioner version)