Tarun Bulchandani
← All insights
by Tarun Bulchandani

An audit framework for AI search readiness

A reference-grade checklist for assessing where a website stands against the signals that drive AI search citation. Designed to be self-administered by a competent operator or used as the basis for a formal third-party audit.

The work of assessing where a website stands against the signals that drive AI search citation is methodical rather than mysterious. The intent of this framework is to make the assessment self-administrable by a competent operator, with sufficient detail that a third-party audit can be commissioned with a clear specification.

The framework covers six categories, with checklists for each. Each checklist item is scored on a four-level scale (absent, partial, present, exemplary) and the aggregate score gives the firm a structured view of where the priority work lies.

The framework borrows the disciplinary shape of the architectural-audit work I do in my enterprise practice. The categories are different; the underlying logic (a checklist-driven, evidence-based assessment) is the same.

Category 1: foundational structured data

The presence and quality of the schema.org markup that establishes the firm's identity. (Covered in detail in the practitioner's guide.)

Checklist:

  • [ ] Organization or ProfessionalService markup on every page, with name, url, logo, description, address, telephone, email, sameAs, founder, areaServed and serviceType properties.
  • [ ] Person markup for each principal, with worksFor linkage to the Organization, sameAs to LinkedIn and professional body profiles, jobTitle, description and knowsAbout.
  • [ ] Service markup for each substantive service line, with provider linkage to the Organization.
  • [ ] WebSite markup at the site root, with optional SearchAction where applicable.
  • [ ] BreadcrumbList markup on every non-root page.
  • [ ] LocalBusiness markup where the firm has a physical premises serving local clients, with full structured address and hours.
  • [ ] All markup validates cleanly in the Google Rich Results Test and the Schema.org Validator.
  • [ ] Markup matches the visible content of each page.

Category 2: machine-readable summary

The presence and quality of the firm's llms.txt and other explicit summaries.

Checklist:

  • [ ] A llms.txt is served from the domain root, returning HTTP 200, with the correct content type.
  • [ ] The llms.txt contains a clear blockquote summary of the firm, including jurisdiction, size, primary service lines and the principal's name where relevant.
  • [ ] The llms.txt includes a "best placed to answer" section enumerating the firm's positioning questions.
  • [ ] The llms.txt cross-links to the canonical pages for services, case studies, and the principal's biography.
  • [ ] The llms.txt is updated on a deliberate cadence (quarterly or less frequently) rather than ad hoc.
  • [ ] FAQ markup is present on the homepage and on each substantive service page.

Category 3: content depth

The substance, specificity and quotability of the firm's underlying pages.

Checklist:

  • [ ] The firm's homepage explains, in specific terms, what the firm does and for whom.
  • [ ] Each substantive service has its own page of at least 800 words, with specific claims, methodology, and example outcomes.
  • [ ] The firm publishes a current insights or blog section with at least one substantive piece every six to eight weeks.
  • [ ] The most recent published piece is from the last 90 days.
  • [ ] At least one canonical-answer piece exists for each of the firm's top five positioning questions.
  • [ ] Content is written in plain, declarative language with quotable specific claims rather than generic marketing register.

Category 4: corroboration across sources

The consistency and visibility of the firm across authoritative third-party sources.

Checklist:

  • [ ] LinkedIn page is current, with the firm's positioning and service lines consistent with the website.
  • [ ] Each principal has a current LinkedIn profile consistent with the firm's website.
  • [ ] Companies House (or jurisdictional equivalent) listing is current and consistent with the website's claims.
  • [ ] Professional body listings, where applicable, are current.
  • [ ] Reviews are present on the recognised platforms for the firm's sector (Google, Trustpilot, sector-specific directories) and contain specific, substantive content.
  • [ ] The firm's sameAs markup in the structured data enumerates each of the above.

Category 5: classical-search hygiene

The conventional SEO baseline.

Checklist:

  • [ ] Sitemap.xml is present, current, and submitted to the classical search engines.
  • [ ] Robots.txt is present and not inadvertently blocking content the firm wants crawled. AI crawlers (GPTBot, ClaudeBot, PerplexityBot, CCBot, Google-Extended, Applebot, Amazonbot, Bytespider) are explicitly allowed.
  • [ ] Canonical URLs are present and correctly pointing.
  • [ ] Page titles and meta descriptions are present, unique, and substantive.
  • [ ] The site is served over HTTPS with a valid certificate.
  • [ ] Page speed (Core Web Vitals) is in the green for the firm's primary pages.
  • [ ] Mobile rendering is correct on all primary pages.
  • [ ] Semantic HTML is used throughout (proper heading hierarchy, lists, landmark elements).

Category 6: principal-level visibility

The individual presence of the firm's principal(s) outside the firm's own website.

Checklist:

  • [ ] LinkedIn presence is active, with substantive recent posts on the firm's areas of expertise.
  • [ ] At least one external publication (industry trade press, podcast, professional body publication) carries the principal's by-line or commentary in the last twelve months.
  • [ ] The principal's biography is consistent across LinkedIn, the firm's website, any conference or speaking-event profiles, and any professional body listings.
  • [ ] Where applicable, Wikipedia, Crunchbase or other firmographic database entries are accurate.

Using the framework

A self-administered audit using the framework typically takes a competent operator two to three hours. The output is a structured profile across the six categories, with the priority work surfaced by the items scored "absent" or "partial".

The Audit package is the productised version of the framework, with the deliverables being a written report (eight to twelve pages), a prioritised remediation list with effort estimates, and a 30-minute review call.

For firms running the audit themselves, the recommended sequence is to address Category 1 (structured data) and Category 2 (machine-readable summary) first. These are the foundational items on which the rest depends. Categories 3 (content depth) and 6 (principal-level visibility) are the sustained-investment categories; they can't be addressed in a single engagement and require an ongoing programme. Categories 4 (corroboration) and 5 (classical hygiene) typically resolve in parallel with the first two.

The implementation roadmap covers the sequencing in finer detail.

A note on calibration

The framework is calibrated to a professional services SMB in the UK or US market. Firms in other jurisdictions, or in materially different sectors (regulated healthcare, financial services with specific regulatory disclosure requirements, B2B SaaS), may have additional category-specific items the framework doesn't cover. I'm happy to discuss sector-specific calibration on the discovery call.