AI Visibility Audit for B2B SaaS: A 20-Point Checklist to Improve Your Visibility

73% of the searches happening globally are exposed to AI generated output.

This is the conclusion we reached while researching the influence of AI Overviews on Google: of the roughly 581 billion searches happening each month across search engines and LLMs, approximately 423 billion expose users to an AI-generated answer.

An AI visibility audit helps you understand whether your business is accessible to AI search systems, benchmark your brand presence and visibility across the core themes you need to own, and assess whether your content answers your customers' questions and is cited in LLMs.

The common issues we have seen across companies we have audited include technical restrictions that block relevant pages from AI crawlers, content that does not cover the use cases buyers are researching, and a lack of brand visibility outside the website.

Without a structured audit, it is difficult to benchmark your current visibility and decide whether the next step should be fixing technical issues, improving content, or building a stronger brand presence.

In this article, we will walk through the three core pillars of an AI visibility audit: technical accessibility, content quality evaluation, and brand recognition analysis. Together, these audits will help you identify where you are falling short and what you need to work on.

AI visibility audit checklist: Summary

Use this checklist to review the main areas covered in the audit.

Technical accessibility

• Review robots.txt and AI crawler permissions
• Check indexing restrictions and blocked URLs
• Verify crawler access in server and CDN logs
• Check JavaScript rendering and HTML accessibility
• Audit site architecture, internal links and sitemaps
• Validate structured data and schema markup

Content and brand authority

• Review heading hierarchy, structure and readability
• Check content relevance and direct answers
• Evaluate ICP, industry and use-case coverage
• Identify buyer-journey content gaps
• Review freshness, factual accuracy and original insights
• Check brand consistency and third-party mentions

AI visibility measurement

• Define ICPs, personas, industries and geographies
• Build the prompt-universe matrix
• Categorize commercial, pain-point and educational prompts
• Assign prompt weights based on intent and relevance
• Track prompts across LLMs and Google AI search
• Measure mentions, citations and recommendations
• Analyze competitor visibility and brand positioning
• Audit citation sources and authority gaps

The complete AI visibility audit checklist

We have divided the audit into three sections: technical accessibility, content readiness and AI visibility measurement. Each section addresses a different part of how AI search systems discover information and generate answers.

Part 1. Technical AI visibility audit checklist

Technical accessibility is the first thing to evaluate. Before improving content for AI search, you need to verify that important website pages can be accessed and that your technical configurations are not unintentionally restricting search crawlers.

1. Check robots.txt and AI crawler access

Start by reviewing your website's robots.txt file. This file contains instructions that tell compliant crawlers which sections of your website they are allowed to access.

Check whether your current configuration permits the crawlers relevant to your AI visibility strategy. These may include OAI-SearchBot, PerplexityBot, Googlebot, Bingbot and other platform-specific agents.

If your website does not have a robots.txt file, create one when you need explicit crawler rules. You can use a general User-agent: * rule or configure individual user agents depending on which crawlers you want to permit.

Make sure important product pages, service pages, blog articles and other public content are not accidentally disallowed. Also distinguish between search-related crawlers and crawlers used for model training, since they may have different purposes.

2. Check indexing restrictions and blocked URLs

Next, review whether the pages you want to appear in search are indexable.

If an important page is not indexed by Google, it may also have limited visibility in AI-generated search results. Google's AI Overviews and AI Mode rely on Google's search infrastructure to discover and surface relevant content. While other LLMs may use different retrieval systems, indexing and crawlability issues can also affect their ability to discover your pages.

Use Google Search Console to identify pages excluded from Google's index and investigate the reasons. Look for unintended noindex directives, incorrect canonical tags, redirects, broken URLs and other indexing restrictions.

Compare the important URLs on your website with your XML sitemap and verify that the sitemap includes the canonical pages you want search engines to discover.

You should also check snippet restrictions that may limit how content can appear in search results or AI search features.

Not every page needs to be indexed. The objective is to make sure commercially important pages and useful informational content are eligible for discovery in both traditional search and AI-powered search experiences.

3. Verify AI crawler access through server and CDN logs

Even when robots.txt allows a crawler, server or CDN configurations may prevent it from accessing the website.

Review your server logs, CDN security settings and firewall rules to identify whether legitimate AI search crawlers are receiving successful responses or being blocked.

Check for bot-management restrictions, rate limits, CAPTCHA challenges and unusual HTTP errors affecting important URLs.

Where possible, verify crawler identities using the provider's published IP ranges or verification methods rather than relying only on the user-agent name.

This helps distinguish between pages that are technically accessible in a browser and pages that AI retrieval systems can actually request.

4. Check JavaScript rendering and HTML content accessibility

Review whether important information on your website is available in the HTML that crawlers can retrieve.

Some websites depend heavily on JavaScript to display product descriptions, pricing details, feature information and other content. Depending on the retrieval system, this information may not always be available if it only appears after client-side rendering or user interaction.

Compare the initial HTML response with the rendered page and identify any important information missing from the initial response.

Prioritize making core product information, page headings, descriptions and answers available as accessible text.

For an example, have a look at Chris Long's post on FigJam, part of Figma, where he found that the FAQ content was not being retrieved by Google or LLM crawlers.

5. Audit website architecture, internal linking and XML sitemaps

Review how your website pages are organized and connected.

Important product, solution, industry and resource pages should be discoverable through logical navigation and relevant internal links.

Check whether important pages are isolated, whether internal links point to broken or redirected URLs, and whether related pages are connected through descriptive anchor text.

Review the XML sitemap for outdated, duplicate, redirected or non-canonical URLs. Make sure it reflects the important indexable pages currently published on the website.

A clear site structure helps search engines discover content and makes it easier to understand the relationships between your products, industries and use cases.

6. Validate structured data and schema markup

Review the structured data implemented across your website.

Depending on the page, relevant schema types may include Organization, Article, BreadcrumbList, Product and SoftwareApplication.

Check whether the markup is valid, whether required properties are included where applicable, and whether the information matches the content visible on the page.

For example, if your product information or pricing has changed, make sure outdated details are not still present in structured data.

Structured data can help supported systems interpret information about your organization and pages. However, adding schema does not guarantee that an AI platform will cite or recommend your brand.

Part 2. Content AI visibility audit checklist

Once the technical audit is complete, the next step is to evaluate whether your content provides the information that AI search users are looking for.

The objective is to identify content that is difficult to understand, poorly structured, outdated or missing information relevant to your customers.

7. Audit content structure, heading hierarchy and readability

Review the structure of your existing content, particularly important landing pages and blog articles.

Check whether pages use a descriptive H1, followed by logically organized H2 and H3 sections. Each subsection should cover a clear topic without unnecessarily repeating information from other sections.

Look for paragraphs that are too long, confusing sentence structures, unnecessary explanations and formatting errors. For example, in this content, we have broken each section into smaller, digestible paragraphs to improve readability for users and retrievability for LLMs.

Use bullet points for lists, numbered steps for processes and tables where comparisons are easier to understand in a structured format.

The goal is to make the content readable for users and easier for automated systems to extract relevant information.

8. Check content relevance and direct answers to user queries

Evaluate whether each page directly addresses the search intent or customer question it is targeting.

For example, if an article focuses on choosing a CRM for healthcare companies, most of its content should address healthcare-specific requirements, relevant product capabilities and the criteria buyers would use to evaluate solutions.

Avoid adding unrelated information simply to increase article length or cover more keywords.

Check whether the main question receives a clear answer early in the relevant section, followed by explanations, examples or supporting evidence where necessary.

A page should provide enough information to satisfy the intended query without forcing readers to search through unrelated sections.

9. Evaluate content coverage across ICPs, industries and use cases

Review whether your website adequately covers the industries, customer profiles and use cases your product is designed for.

For B2B SaaS companies, this may include dedicated information about industry-specific requirements, buyer responsibilities, common operational problems, integrations and product applications.

Compare your existing content against your ideal customer profiles and identify areas where the product's relevance is not clearly explained.

Avoid creating large amounts of generic content around topics that have little connection to your product or target customers.

Content coverage should be based on the questions your actual buyers are likely to ask.

10. Identify content gaps across the buyer journey

Map your existing content to the questions customers ask during different stages of their buying journey.

Early-stage buyers may be researching a problem or learning about possible approaches. Later-stage buyers may compare products, evaluate alternatives, investigate implementation requirements or look for pricing information.

Review whether your existing pages answer these questions adequately.

For each important topic, identify whether an existing article needs additional information, whether multiple overlapping articles should be consolidated, or whether a new page is required.

Prioritize gaps that are relevant to your ICP and have a clear connection to the product or buying decision.

11. Audit content freshness, accuracy and original insights

Check whether your existing content is accurate and reflects current information.

Review outdated statistics, old product descriptions, unsupported claims, broken references and content that no longer matches your product's capabilities.

Where possible, add original information such as internal research, customer examples, product data, practical observations and specific use cases.

Content should be updated when the underlying information changes or when important details are missing, rather than simply changing the publication date.

For commercially important pages, verify that product features, integrations, pricing and other decision-related information are current.

12. Evaluate brand authority, entity consistency and third-party mentions

Review how your brand is represented across your website and external platforms.

Your company description, product category, positioning, pricing information and major product capabilities should remain consistent across relevant sources.

Check official profiles, software directories, review platforms, partner websites and third-party articles where your company is mentioned.

For example, if your company has changed its product positioning or pricing, an older directory listing may still describe the previous offering. Identify these inconsistencies and update the listings you control or contact the relevant publishers.

Also identify credible third-party platforms where competitors are mentioned but your brand is absent.

This can help you plan relevant directory listings, independent reviews, industry coverage and other forms of third-party brand presence.

Part 3. AI visibility measurement and performance audit

After reviewing technical accessibility and content readiness, the next step is to measure how your brand actually performs in AI-generated answers.

This requires building a relevant prompt set, tracking responses across different platforms and comparing your brand's visibility with competitors.

13. Define your ICP, buyer personas, industries and target geographies

Start by defining the scope of your AI visibility measurement.

Identify the products or solutions you want to evaluate, your ideal customer profiles, buyer personas, priority industries and target geographic markets.

Also establish the product categories and competitor groups relevant to those customers.

For example, a SaaS company targeting enterprise healthcare organizations in the United States should not rely entirely on generic software prompts that apply to small businesses across every industry.

A defined scope helps you measure visibility in the buying situations that actually matter to your company.

14. Build a prompt-universe matrix based on ICPs and buyer contexts

Use the defined ICPs and buyer contexts to build a structured prompt-universe matrix.

The matrix should combine relevant dimensions such as industry, buyer persona, geography, company size, business problem and product category.

For example:

Use these combinations to create natural-language prompts that reflect how different customers research problems and evaluate solutions.

The objective is to build a representative prompt set rather than generating hundreds of repetitive questions.

15. Categorize prompts into commercial, pain-point and educational themes

Once the prompt universe is created, divide the prompts into three main themes.

Commercial prompts focus on product discovery, comparisons, alternatives and vendor selection. These are generally associated with bottom-of-funnel buying activity.

Pain-point prompts focus on specific operational or business problems that the product helps solve. Depending on the wording, they may reflect middle- or bottom-of-funnel intent.

Educational prompts focus on explanations, industry concepts, processes and general information. These are typically associated with top- and middle-of-funnel research.

For example:

This classification makes it easier to compare visibility across different types of customer questions.

16. Assign weights to prompts based on commercial intent and relevance

Not every prompt has the same commercial value.

A prompt asking for the best software vendors in a category may be more relevant to an immediate buying decision than a general educational question.

Assign weights based on commercial intent, ICP relevance and strategic importance.

For example, you could use a three-point scale, with high-intent commercial prompts receiving a weight of three, relevant pain-point prompts receiving two and broad educational prompts receiving one.

These weights should be treated as an internal prioritization method rather than an industry-standard AI visibility score.

Track high-intent performance separately so that strong educational visibility does not hide weak visibility in product recommendations.

17. Track prompts across LLMs and Google AI search experiences

Run your selected prompts across the AI platforms relevant to your target audience.

These may include ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews and AI Mode.

Google AI search experiences should be evaluated separately from standalone LLM platforms, since their interfaces, triggering behavior and available measurement data differ.

For each prompt, record the platform, exact question, date, relevant geographic or language context and generated answer.

Repeat important prompts under consistent conditions to account for variations in AI responses.

The goal is to establish a reliable comparison across platforms rather than relying on a single response.

18. Measure brand mentions, citations and recommendation rates

Measure three separate types of AI visibility.

A brand mention occurs when your company is named in an AI-generated answer.

A citation occurs when the platform links to or references your website or content as a supporting source.

A recommendation occurs when the platform suggests your product or company as a suitable option for the user's question.

These outcomes should not be treated as equivalent.

For example, an AI answer may cite your company's research while recommending a competitor's software.

Record how frequently each outcome occurs across the relevant prompt set, and calculate recommendation rates separately for commercial prompts.

19. Analyze competitor visibility and brand positioning in AI answers

Compare your brand's performance with competitors across the same prompts.

Identify which competitors are mentioned or recommended most frequently and which industries, use cases or buyer contexts they appear to perform well in.

Review how AI platforms describe your brand, including its category, target customers, capabilities and differentiators.

Look for incorrect descriptions, outdated positioning and cases where a competitor is recommended for a use case your product also supports.

This analysis helps identify whether your visibility gap is related to content coverage, brand positioning, supporting evidence or actual product fit.

20. Audit citation sources and identify authority gaps

Review the websites and URLs cited in AI-generated answers, particularly for commercial prompts where competitors are recommended.

Identify whether those sources include competitor websites, independent listicles, software directories, review platforms, industry publications or community discussions.

Look for sources that repeatedly appear across relevant prompts but do not mention your brand.

For example, if a third-party comparison article is frequently cited when users ask for software recommendations, review whether your product belongs in that comparison and whether the publisher can be approached with accurate information.

Also identify opportunities to contribute useful answers on relevant community platforms such as Reddit or Quora, following their rules and avoiding promotional spam.

The purpose is to understand which sources support AI answers and where your brand may need stronger public evidence.

How to prioritize findings from an AI visibility audit

After completing the technical, content and measurement audits, you should have a clearer picture of where your brand is visible, where competitors appear instead and what may be limiting your presence.

Use the prompt weights and repeated measurements to prioritize commercially relevant gaps. A missing recommendation for a high-intent buyer prompt may deserve more attention than an absence from a broad educational question.

Group the findings into the following actions:

1. Resolve technical accessibility issues. Fix blocked URLs, indexing restrictions, rendering problems and crawler-access issues affecting important content.
2. Create content for missing buyer questions. Identify commercially relevant prompts where your brand has limited visibility and your website lacks a page that adequately answers the question.
3. Improve existing content. Update pages that already cover the topic but lack specific information, relevant use cases, comparisons or supporting evidence.
4. Increase relevant third-party mentions. Review directories, industry articles, software comparisons and other sources that are frequently cited. Identify legitimate opportunities to correct listings, contribute information or earn inclusion.
5. Build stronger participation in relevant communities. Where buyer questions appear on Reddit, Quora or industry forums, contribute useful answers and relevant expertise rather than publishing promotional mentions.
6. Measure changes against the original baseline. Continue tracking the same priority prompts and compare mentions, citations, recommendations and competitor presence over time.

Not every visibility gap requires a new article. Sometimes an existing page needs better information. In other cases, the missing evidence is on third-party websites or the product is simply not the right fit for that particular query.

The audit should help distinguish between these situations before resources are committed.

If you're unsure where your website stands in AI search, ScaleTheory can help you identify the gaps. From technical crawler accessibility and content readiness to prompt tracking and competitor visibility, our team can conduct an AI visibility audit and recommend the improvements that matter most to your business. Get an expert AI visibility audit from ScaleTheory

Frequently asked questions

How is an AI visibility audit different from an SEO audit?

A traditional SEO audit examines how well a website can be crawled, indexed and ranked in search engines. An AI visibility audit also evaluates how a brand is mentioned, cited, described and recommended in AI-generated answers, including how it performs against competitors for relevant buyer questions.

Which AI crawlers should you check during an audit?

Check the crawlers relevant to your target platforms, including OAI-SearchBot, PerplexityBot, Googlebot and Bingbot, along with applicable Anthropic crawlers. Review each provider's documentation because crawlers used for search, training and user-initiated retrieval may have different functions.

How many prompts should you track for AI visibility?

For an initial audit, 20–30 carefully selected prompts can provide a useful starting point. Build them around your ICPs, industries, buyer personas and commercial priorities. Larger companies may need separate prompt sets for different products, markets or customer segments.

How often should you conduct an AI visibility audit?

A full technical and content audit can be conducted quarterly or after major website changes. High-priority AI search prompts can be monitored more frequently, such as weekly or monthly, depending on your resources and how quickly the market changes. Keep the core prompt set consistent when comparing results.

Does schema markup or llms.txt improve AI visibility?

Structured data can help supported systems understand information about your website and business, but it does not guarantee AI citations or recommendations. The llms.txt file is an optional convention and is not a universal requirement for appearing in AI-generated answers.