AI SEO Audit: How to Run an AI Visibility Check on Your Site

15+ years in local marketing; Google Ads certified; Shopify Partner.

Here's a scenario I see constantly. A business owner gets their monthly SEO report. It's got ranking screenshots, keyword position tables, a few green arrows. Looks fine. But when they ask "how many leads came from search this month?" , silence. Or worse, a pivot to impressions.

That's a traditional SEO audit. And it's missing half the picture now.

An AI SEO audit is different. It checks whether your site is visible in the places people are actually searching in 2026 , not just Google's blue links, but Google AI Overviews, ChatGPT, Perplexity, and whatever else is eating into your organic traffic. If you haven't checked your site against these channels yet, this guide walks you through exactly how to do it.

What this article covers: the specific checks, the order to run them, and what to do when something fails. What it doesn't cover: the broader question of how AI is changing SEO strategy overall. For that, our complete AI SEO playbook is where to start.

Why Your Old SEO Audit Misses Half the Picture

Google AI Overviews now appear on more than 39% of informational queries, per 2024 BrightEdge research. That means a huge chunk of searches that used to send traffic to your site now get answered before anyone clicks anything.

ChatGPT hit 800 million weekly active users in early 2026. Perplexity is growing fast. Google AI Mode launched in May 2025 and is a separate experience from traditional search entirely. (For a breakdown of how these differ, see Google AI Mode: What It Means for Canadian SMBs.)

The practical effect: your site can rank #1 in Google's traditional results and still be invisible to a big portion of your potential customers. Because they're asking ChatGPT, not typing a query and scrolling.

A traditional SEO audit checks technical health, keyword rankings, and backlinks. All still important. But it won't tell you whether an AI model knows your business exists. That's the gap an AI SEO audit fills.

In my experience, most Canadian SMB sites fail at least two or three of the checks below. Not because they're doing anything wrong, exactly , just because nobody told them the game had changed.

The 5 Checks That Make Up an AI SEO Audit

Think of this as five layers. Each one builds on the last. You can run most of these yourself in a couple of hours.

Check 1: AI Crawlability (Can the Bots Even Read Your Site?)

Before any AI model can mention your business, its crawler has to be able to read your site. There are a few crawlers worth knowing about: GPTBot (OpenAI), ClaudeBot (Anthropic), and PerplexityBot.

Open your robots.txt file. It's at yourdomain.com/robots.txt. Look for any lines that say Disallow next to these bot names. If you're blocking them, AI models can't index your content. That's a hard stop.

A lot of sites block these crawlers by accident , either because a developer added a blanket disallow rule years ago, or because a security plugin got overzealous. For a full breakdown of what to do with each bot, see our guides on GPTBot and ClaudeBot.

Also check whether you have an llms.txt file. This is a newer convention that helps AI models understand what your site is about and how to navigate it. Most SMB sites don't have one yet. If you want to set one up, our llms.txt setup guide walks through it.

Check 2: Structured Data (Are You Speaking AI's Language?)

AI models don't read your site the way a human does. They parse structured signals. Schema markup , that's code you add to your site that labels what things are , makes it dramatically easier for AI to understand your business, your services, your location, and your credibility.

Run your homepage and a key service page through Google's Rich Results Test (free, at search.google.com/test/rich-results). Look for:

LocalBusiness or Organization schema
FAQPage schema on any FAQ content
Service schema on your service pages
Review or AggregateRating schema if you have testimonials

If you're missing these, that's a straightforward fix. For more detail on what schema actually matters for AI search specifically, see Schema Markup for AI Search: What Actually Matters.

Check 3: Content Answerability (Does Your Site Actually Answer Questions?)

Here's the thing about AI search. It's built on questions. Someone types "who's the best plumber in Saskatoon" or "what does a dental implant cost in Regina" into ChatGPT, and the model goes looking for content that directly answers that question.

If your site only has service pages that say "We offer premium plumbing services across Saskatchewan," you're invisible in that query. You need content that actually answers the questions your customers are asking.

Pull up your Google Search Console (free, at search.google.com/search-console). Go to the Performance report. Filter by query type and look for question-format queries , anything starting with "how," "what," "why," "best," "cost of." These are the queries where AI Overviews and AI search are most active.

Then check: does your site have a page that directly answers each of those questions? Not a page that loosely relates to the topic. A page that answers it, clearly, in the first 200 words.

Typically, sites that have FAQ sections on service pages or standalone blog posts answering specific questions tend to show up in AI-generated answers within 4-8 weeks of publishing. Sites that don't have this content almost never appear. That's the pattern I keep seeing.

Check 4: Brand Presence in AI Answers (Does AI Know You Exist?)

This is the check most people skip, and it's probably the most important one.

Open ChatGPT, Perplexity, and Google (if AI Overviews are active in your account). Run a few test queries:

"[Your service] in [your city]" , e.g., "family dentist in Regina"
"[Your business name]" , direct brand query
"[Problem your business solves]" , e.g., "how much does a dental implant cost in Saskatchewan"

See what comes back. Does your business get mentioned? Do your competitors show up and you don't? What sources are being cited?

This is your AI visibility baseline. Write it down. You'll want to track it month over month. For a structured way to do this, our guide on AI search visibility tracking covers the full process.

If you're not showing up and your competitors are, that's usually a content authority problem. The AI models are pulling from sources they trust , often review sites, local directories, industry publications, and well-structured service pages. If your content isn't in those places, you're not in the mix.

For a deeper look at how to actually get cited by ChatGPT and Perplexity, see How to Earn AI Citations from ChatGPT and Perplexity.

Check 5: Technical Foundations (The Stuff That Still Matters)

AI search didn't make traditional technical SEO irrelevant. It just made it table stakes instead of a differentiator.

Run your site through Google PageSpeed Insights (free, at pagespeed.web.dev). You want a score above 70 on mobile. Below that, Google's own crawlers deprioritize your content , and if Google's crawlers are deprioritizing you, AI models that pull from Google's index will too.

Check your Core Web Vitals in Search Console. Look at the "Experience" report. Red or orange signals there are worth fixing before anything else.

Also check: is your site indexed at all? Go to Google and type site:yourdomain.com. If fewer pages come back than you expect, you have an indexing problem that needs to be fixed before any AI optimization will matter.

How to Run This Audit Week by Week

Here's how I'd sequence this if you're doing it yourself.

Week 1: Crawlability and technical foundations. Check robots.txt for bot blocks. Run PageSpeed Insights on your homepage and top 3 service pages. Pull your Core Web Vitals report from Search Console. Fix anything that's actively broken before moving on.

Week 2: Structured data. Run your key pages through the Rich Results Test. Document what schema is present and what's missing. If you're comfortable editing your site's backend, add LocalBusiness and FAQPage schema first , those two tend to have the biggest impact for SMBs. If you're not comfortable with code, this is a reasonable thing to hand off to a developer for a few hours of work.

Week 3: Content answerability. Pull your top 20 queries from Search Console. For each question-format query, check whether your site has a page that directly answers it. Make a list of content gaps. Prioritize by search volume. Start writing or briefing the highest-priority pieces.

Week 4: AI visibility baseline. Run your test queries in ChatGPT, Perplexity, and Google. Screenshot the results. Note which competitors appear. This is your starting point , you'll compare against it in 60 and 90 days.

Month 2 onward: Content and citations. Start publishing the content you identified in Week 3. Build out your presence on the directories and sources AI models cite , Google Business Profile, industry directories, review platforms. Track your AI visibility monthly.

This isn't a one-time project. It's more like a quarterly check-up. The AI search landscape is shifting fast enough that what worked six months ago might not be enough now.

What to Do When Your Audit Turns Up Problems

Most sites I look at have the same handful of issues.

If you're blocking AI crawlers: Fix your robots.txt immediately. This is the easiest win on the list.

If you're missing schema: Add it. If you're on WordPress, plugins like Yoast or RankMath handle the basics. If you're on a custom build, you'll need a developer. Budget a few hundred dollars for a one-time implementation.

If your content isn't answering questions: This is the biggest lift, but also the highest-value work. Plan for 3-6 months of consistent content production before you see meaningful changes in AI visibility. There's no shortcut here. For help thinking through what content to produce, our AI content writing guide for SMBs is worth reading.

If you're not showing up in AI answers: Start with your Google Business Profile. Make sure it's complete, accurate, and actively collecting reviews. Then look at whether your business is listed in the directories that AI models tend to cite for your industry. This is a slower process , building the kind of authority AI models trust takes time and consistent effort.

If your technical scores are poor: Prioritize mobile speed above everything else. A slow mobile site is a problem for traditional SEO and AI search alike.

One thing I want to be honest about: if your audit turns up major issues across multiple areas, trying to fix everything at once usually means fixing nothing well. Pick the one or two highest-priority items and do those properly before moving to the next layer.

When to DIY vs When to Hire

Most of the checks in this audit are genuinely doable without an agency. Google's free tools (Search Console, PageSpeed Insights, Rich Results Test) cover a lot of ground. If you're reasonably comfortable in your website's backend, you can handle a lot of the fixes too.

Where it gets harder to DIY:

If your robots.txt issues are tangled up with a complex server configuration
If your schema implementation keeps breaking
If your content gaps are significant and you don't have internal bandwidth to produce good content
If you need to track AI visibility across multiple products or locations

That's the piece where a second set of eyes tends to pay for itself. Not because the work is mysterious, but because it's time-consuming and easy to do halfway.

If you're evaluating agencies on this front, our AI marketing agency guide covers what to ask and what to pay. And if you want to understand what a full AI readiness audit looks like at the agency level, here's what that typically includes.

For the broader strategic question of how AI fits into your marketing overall, our AI for marketing guide is the right place to start.