The AIO Readiness Audit: How to Assess Your Brand’s AI Search Visibility in 90 Minutes

The most common mistake I see marketing teams make when they start thinking about AIO is skipping the audit entirely. They read about AI Optimization, understand that it matters, and jump straight into creating new content, adding schema markup, or optimizing their service pages for AI citation. Some of that work may be correct. But without a baseline, there is no way to know which of the five layers of AIO actually has the gap that is costing them visibility, and which layers are already working well enough that effort there is wasted.

The Semrush AI Visibility Index 2026, which analyzed 126 million real user prompts across ChatGPT, Gemini, Google AI Mode, and Google AI Overviews, found that only 9 percent of marketing teams can measure all the metrics that matter for AI search visibility. Forty-five percent cannot properly measure AI visibility at all. You cannot close a gap you have not measured.

This audit is designed to be completed by any marketing team, with no specialized tools required beyond GA4, Google Search Console, and a browser, in approximately 90 minutes. It covers all five layers of the Infidigit AI Discovery Framework: Foundation, Semantic Architecture, AI Extractability, Entity and Citation Authority, and Measurement. Each step produces a specific finding that drives a specific action. The goal is not a theoretical assessment but a prioritized action list your team can execute against the week after you complete it.

Before you begin, open a blank document to record your findings as you go. At the end of each step, score yourself using the framework provided. Your total score will tell you which layer to prioritize first.

Layer 1: Foundation Audit

The foundation layer determines whether AI systems can physically access, crawl, and read your website. Most content and schema work is irrelevant if this layer has failures, because AI crawlers that cannot reach your content cannot index or cite it regardless of how well-structured it is.

Step 1  AI Crawler Access Check  |  15 minutes

Go to your website’s robots.txt file by adding /robots.txt to your domain in a browser. Read every User-agent entry and every Disallow directive.

You are looking for: GPTBot (OpenAI), OAI-SearchBot (OpenAI), ClaudeBot (Anthropic), Claude-SearchBot (Anthropic), PerplexityBot, GoogleOther (used for AI Overview training and retrieval), Googlebot-Extended.

A wildcard disallow, written as User-agent: * followed by Disallow: /, blocks all crawlers that are not explicitly allowed. If this exists and the AI crawlers above are not listed with explicit Allow directives above the wildcard rule, every one of them is blocked from your site. This is the most common and most damaging Layer 1 failure we encounter.

Also check your CDN or hosting provider’s bot management settings. Cloudflare, Fastly, and similar services have security features that can block AI crawlers at the infrastructure level, separate from your robots.txt. A site with a clean robots.txt can still be blocking GPTBot at Cloudflare without the team knowing.

Finally, load three of your most important pages in a browser with JavaScript disabled (in Chrome: Settings, More tools, Developer tools, Rendering tab, check Emulate CSS media type, then under Rendering click the dropdown at bottom and select no JavaScript). The content that renders is what most AI crawlers can access. If key paragraphs, product descriptions, or FAQ content disappears when JavaScript is disabled, AI crawlers are not seeing that content.

Step 2  llms.txt and Organization Schema Check  |  10 minutes

Navigate to yourdomain.com/llms.txt. If the file exists, read it. Does it accurately describe what your brand does today? Does it list your most important pages? If the description is outdated, or if the file does not exist, this is a gap.

Then view the source of your homepage (right-click, View Page Source, search for ‘Organization’). Check whether Organization schema exists as a JSON-LD block. If it does, verify it includes: name, url, logo, description, foundingDate, and at minimum one sameAs reference pointing to a verified external profile such as LinkedIn.

The sameAs connection is what links your website entity to your external entity footprint. Without it, an AI system that separately encounters your LinkedIn page and your website has no structured signal confirming they are the same brand.

Layer 2: Semantic Architecture Audit

The semantic architecture layer determines whether your content is organized so that AI systems can understand what your brand’s subject domain is, what sub-topics you cover, and how your content connects. A site with strong crawlability but no semantic architecture will be accessed by AI crawlers but cited inconsistently, because the topic relationships are unclear.

Step 3  Topical Map Existence Check  |  10 minutes

Open a spreadsheet. List every service, product, or topic your brand covers. Now map these into a hierarchy: broad parent topics, sub-topics, and micro-topics. Do you have at least one page for each level of that hierarchy?

The most common gap here is a content estate that consists entirely of commercial pages, service pages, and location pages, with no educational or informational content sitting beneath them to signal topical depth. A site with 30 service pages and zero educational blog or insights content has a flat architecture that AI systems interpret as a directory, not an authority.

Check your blog or insights section. Count the number of posts. Are they connected to your service pages through contextual internal links, or do they sit as isolated pages with no semantic relationship to the commercial content? The goal is a hub-and-spoke architecture where every piece of informational content links to and from the relevant service page.

Step 4  H2 Heading Structure Check  |  10 minutes

Open three of your most important pages, ideally your primary service page, your most-visited blog post, and your homepage. Read every H2 heading on each page.

Count how many H2 headings are written as questions versus statements. A statement heading like ‘Why LLM SEO Matters’ is a label. A question heading like ‘What is LLM SEO and how does it improve AI search visibility?’ is both a question a buyer would actually type into an AI assistant and a structural signal to AI retrieval systems that this section answers a specific query.

Then read the first two to three sentences directly under each H2. Do those sentences give a complete, standalone answer to the heading question? If you pulled just that heading and those first sentences out of the page, would they make sense as an independent answer? If yes, that section is extractable. If no, it is not.

This is the extractable unit principle, and it is the most consistently underimplemented element in content we audit. The global ecommerce platform that achieved 1,139x growth in AI-generated traffic had its content restructured around exactly this principle: every section independently answerable, every opening sentence directly responsive to the heading question.

Layer 3: AI Extractability Audit

AI extractability determines whether AI systems can retrieve clean, accurate, independently meaningful answers from your content. This layer focuses on structured data and content precision, not just content volume. Improving this extractability is also a core part of LLM SEO, where content is structured to be easier for AI systems to understand, retrieve and reference.

Step 5  Schema Markup Audit  |  15 minutes

Go to Google’s Rich Results Test tool (search.google.com/test/rich-results). Test three URLs: your homepage, your primary service page, and your most-visited piece of content.

For each page, record which schema types are detected. The minimum viable schema set for AIO is:

• Organization schema on the homepage with name, url, logo, description, and sameAs references
• Article or BlogPosting schema on all authored content with author (linking to Person schema), datePublished, and dateModified
• Person schema for Kaushal on all authored content pages, linking to LinkedIn
• FAQ schema on any page that contains question-and-answer sections
• HowTo schema on any page that describes a step-by-step process

The B2B ecommerce platform that achieved 245x improvement in AI Overview keyword presence had schema implementation as one of the core solutions: Product, Category, and Blog schema implemented to increase eligibility for rich results. The ICICI Prudential engagement that produced 39x LLM traffic growth included Organization, Product, FAQ, and Dataset schema as part of the AIO framework. Schema is not a secondary consideration; it is what makes your content claims machine-readable.

Note every schema type that should exist on each page but does not. This becomes your schema implementation priority list.

Step 6  Content Precision Check  |  10 minutes

Open your primary service page or homepage. Read the first 200 words carefully.

Ask yourself three questions about what you just read:

First: does it contain a specific, verifiable claim with a number attached? Not ‘we deliver exceptional results’ but ‘Infidigit’s LLM SEO engagements have produced between 37x and 1,139x growth in AI-referred sessions across 200-plus client engagements.’ The specific claim is citable. The general claim is not.

Second: does it use Entity-Attribute-Value structure? Does it name an entity (your brand or a client), state an attribute (the service delivered or the outcome achieved), and give a specific verifiable value (the number, the timeframe, the market)? EAV-structured content is what AI systems extract and cite. Generic marketing language is what they skip.

Third: is there anything in the first 200 words that an AI retrieval system could extract as a standalone answer to a question a buyer might ask? If someone asked ChatGPT ‘what does Infidigit do for enterprise brands’ and the AI retrieved only your first 200 words, would it have enough to give a useful, specific answer?

This precision check matters because AI systems prioritize dense, specific, factual content over marketing copy. The Dun and Bradstreet engagement that produced 57x LLM traffic growth and 1,500-plus AI Overview placements started with identifying exactly this gap: content that was SEO-optimized but not structured for AI extraction.

Layer 4: Entity and Citation Authority Audit

The entity and citation authority layer determines whether AI systems can corroborate your brand claims through independent external sources, and whether those sources describe you consistently. This layer is addressed in full in the companion article in this series on off-site entity clarity. Here, the audit focuses on quickly identifying where the most significant gaps are.

Step 7  Off-Site Entity Consistency Check  |  10 minutes

Open five tabs and navigate to: your LinkedIn company page, the top result when you search your brand name on Google (usually a Wikipedia page or Crunchbase if either exists), your Clutch profile, the most recent press article about your brand, and the top Reddit thread mentioning your brand name if one exists.

Read the description of your brand on each source. Answer these questions:

• Does each source describe the same core service or product offering in consistent language?
• Does each source use the same key terminology? If your website says AIO, does your LinkedIn say AI search optimization or digital marketing?
• Does each source describe your current positioning, or does any of them reflect an older version of what you do?
• Is your brand missing from any of these sources entirely?

The Semrush 2026 data shows that the brands with the most stable, cross-platform AI visibility have one characteristic in common: their third-party sources describe them consistently. Patagonia’s AI Visibility score held at 79 or 80 across all four major AI platforms across every month of the measurement window because specialist outdoor review sites consistently describe the brand with the same vocabulary. Consistency is built upstream of AI, in the ecosystem of sources AI draws from.

Layer 5: Measurement Audit

Step 8  AI Traffic Measurement Check  |  10 minutes

Open GA4 and navigate to Reports, Life cycle, Acquisition, User acquisition. Look at the Default Channel Group column. Do you see an AI Assistant channel? If yes, click on it and check the session volume over the last 90 days. This channel was added natively by Google on May 13, 2026 and should appear without any configuration.

If you do not see an AI Assistant channel, your GA4 property may not have updated to the new default channel group definitions. Check whether your sessions from chatgpt.com, perplexity.ai, claude.ai, and gemini.google.com are appearing in Referral or Direct traffic instead.

Next, open Google Search Console and navigate to Performance, Search results. Check whether you have a Generative AI tab available (launched June 3, 2026). If present, click it and record the total impressions over the last 28 days. This is your AI Overview impression count: the number of times your pages appeared inside Google’s AI features.

Finally, open Search Console’s regular Performance report and filter for your brand name as a query. Note the click volume trend over the last three months. Branded search lift, the increase in direct searches for your brand name, is the downstream signal of AI citation activity that does not produce a tracked click.

Step 9  Citation Rate Baseline Run  |  10 minutes

This is the most important single step in the audit because it produces the only direct, empirical measure of your current AI visibility: how often does your brand actually appear in AI-generated answers?

Open ChatGPT, Perplexity, and Gemini in three browser tabs. Run these five prompts in each, typing them exactly as written:

• Who are the best [your service category] agencies or companies for [your target customer segment]?
• What are the top options for [the primary problem your brand solves]?
• What is [your brand name] known for?
• What do people say about [your brand name]?
• Which [your service category] company would you recommend for an enterprise brand?

Record for each prompt and each platform: whether your brand appears, what it says if it does, whether the description is accurate, and which competitor brands appear alongside or instead of you.

This prompt-based audit is exactly how Infidigit establishes the baseline for every AIO engagement. ICICI Prudential started with fewer than 200 AI Overview keyword placements before the engagement. After baseline measurement, entity optimization, content restructuring, and schema implementation, they reached 1,093 AI Overview placements and a 39x increase in LLM platform sessions. The baseline is what makes that progress visible and attributable. Without it, you cannot distinguish between AI visibility that was already there and AI visibility that your optimization work produced.

Scoring: What Your Result Means and What to Do Next

Total your scores across all nine steps. The maximum is 27 points.

What the Audit Tells You Beyond the Score

The numerical score is a useful triage tool, but the most valuable output of the audit is the specific list of findings from each step. Each gap you identified maps to a specific task:

• A Step 1 finding about blocked AI crawlers maps to a robots.txt or CDN configuration task assigned to a developer, with a target completion date of this week.
• A Step 2 finding about missing Organization schema maps to a JSON-LD implementation task, typically under an hour of dev time once the content is written.
• A Step 3 finding about no topical map maps to a content planning session that produces a documented hierarchy, followed by an editorial calendar that builds out the gaps.
• A Step 4 finding about statement-based H2s maps to a content editor pass across the three to five highest-priority pages, restructuring headings and adding direct opening answers.
• A Step 5 finding about missing FAQ schema maps to a structured data implementation across every page that has a Q-and-A section, which may be an Elementor or WordPress plugin task rather than a custom dev task.
• A Step 6 finding about generic content in the first 200 words maps to a rewriting task for the opening paragraphs of each priority page, replacing marketing language with specific, EAV-structured claims.
• A Step 7 finding about inconsistent external descriptions maps to the off-site entity audit and correction process described in Article 6 of this series.
• A Step 8 finding about missing AI traffic measurement maps to a GA4 channel group setup task and a Search Console check, both completable in under an hour.
• A Step 9 finding about low citation rate baseline maps to a prioritization decision: which of the upstream layers (1 through 4) is most likely to be the root cause, and which layer-specific task should come first.

The pattern across the AIO engagements Infidigit has run is consistent: brands that treat the audit as the starting point and work through each layer in sequence, rather than jumping to content or schema work before the foundation is clean, produce faster and more durable results. The B2B ecommerce platform that achieved 245x AI Overview keyword improvement started with a complete SEO and AIO audit to analyze technical barriers, content gaps, and optimization opportunities before any content or schema work was touched. The audit is not preparation for the work. It is the first deliverable.

The 10-Question AIO Maturity Self-Check

The Semrush 2026 AI Visibility Index includes a 10-question maturity self-assessment that I find useful as a complement to the layered audit above. It focuses on measurement and operational maturity rather than the technical and content specifics the layered audit covers. Score one point for each yes.

1. Is your AI Visibility spread across your key platforms within 10 points (meaning you are not heavily concentrated on one platform)?
2. Are at least 30 percent of your AI citations from your own domain, rather than entirely from third-party sources?
3. Have your brand’s AI mentions grown in the past three months?
4. Does your brand have a single named internal owner of AI search?
5. Do you track mentions and citations as separate metrics?
6. Have you mapped your competitive co-occurrence set, meaning the brands that appear alongside yours in AI responses?
7. Do you measure AI visibility per platform separately, not as a single aggregated score?
8. Do you appear among your industry’s top brands on multiple platforms, not just one?
9. Are authoritative third-party sources cited alongside your own content in AI responses about your brand?
10. Is your brand described with consistent language across the sources AI quotes from?

One More Thing Before You Start

The purpose of this audit is not to produce a score. It is to produce an action list with a clear priority order. When you finish, you should have: a robots.txt status confirmed, a schema implementation list, a list of H2 headings to rewrite, a list of external sources to audit for consistency, a GA4 and Search Console measurement setup status, and a citation baseline from the five prompts.

That list is the brief that goes to your team the same week. Not next quarter. The Semrush data found that 81 percent of teams with fully integrated SEO and AI search execution report more traffic or leads from AI platforms. That integration starts with knowing where you actually are, which is what the audit gives you.

If after completing the audit your score suggests that the gaps are beyond what your internal team can address quickly, the case studies listed at the end of this article show what an external AIO engagement can produce and at what pace. The ICICI Prudential engagement went from under 200 AI Overview placements to 1,093 within the engagement timeline. The global ecommerce platform went from near-zero AI traffic to a 1,139x increase in LLM sessions over the project period. Those are not exceptional results. They are what happens when the five-layer framework is applied systematically, starting with the audit.

Kaushal Thakkar

Founder & CEO, Infidigit

Kaushal has spent over a decade building organic growth programs for enterprise and mid-market brands, advising CMOs and marketing leaders across 200+ brands. He now leads Infidigit’s shift from traditional SEO into AI-driven discovery.

• Follow the Expert