How often should you run a technical SEO audit?

Run a full audit at the start of an engagement and at least quarterly, with automated checks weekly or daily for issues that cause sudden damage, such as robots.txt changes, noindex tags and server errors.

What is the most important part of a technical SEO audit?

Crawlability and indexation. If search engines and AI crawlers cannot reach and index the important pages, nothing else in the audit matters.

What is new in technical SEO audits for AI search?

Two checks: whether AI search and user-triggered crawlers are allowed in robots.txt and firewall rules, and whether important content is present in the server-rendered HTML, because most AI crawlers do not run JavaScript.

How long does a technical SEO audit take?

For a small site, a day or two. For large e-commerce or enterprise sites, one to three weeks, mostly spent on analysis and prioritisation rather than running the crawl.

What tools do you need for a technical SEO audit?

A crawler that can render JavaScript and crawl without it, Google Search Console, a page speed tool using field data, a structured data validator, and access to server or CDN logs for crawler analysis.

Should a technical SEO audit include server logs?

Yes, where you can get them. Logs show which pages search engines and AI crawlers actually request, how often, and what errors they hit, which crawler tools and Search Console cannot fully show. Even 30 days of logs often reveals blocked crawlers, wasted crawl on parameter URLs and important pages that are never visited.

Technical SEO Audit Checklist for 2026 (With AI Search Checks)

A technical SEO audit answers one question: can search engines and AI systems reach, read, understand and index the pages that matter? In 2026 that question has a second audience. Googlebot still matters most for organic traffic, but OpenAI's, Anthropic's and Perplexity's crawlers now decide whether a site can be cited in AI answers, and they behave differently.

This checklist covers both. Work through it in order, because each section depends on the ones before it. Record every issue with the affected URLs, its likely impact and the fix, then prioritise at the end.

1. Crawlability

Crawlers must be able to request pages before anything else can happen.

robots.txt is reachable at the root and returns a 200 status.
No important sections are disallowed by mistake, including CSS and JavaScript files needed for rendering.
XML sitemaps exist, are referenced in robots.txt, list only canonical, indexable URLs that return 200, and are submitted in Search Console.
Server responds reliably. No significant rate of 5xx errors in Search Console's crawl stats or in server logs.
Firewall and CDN bot rules do not challenge or block legitimate search crawlers.
Crawl budget is not wasted on infinite parameter combinations, faceted navigation or calendar pages.

2. AI crawler access

This section is new for most audits, and it is often where the biggest AI visibility problems hide.

AI search crawlers are allowed: OAI-SearchBot, Claude-SearchBot and PerplexityBot, unless the client has chosen otherwise.
User-triggered agents are allowed: ChatGPT-User, Claude-User and Perplexity-User.
Training crawlers reflect the client's decision: GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, CCBot and others are allowed or blocked deliberately, and the decision is documented.
Googlebot is not blocked in an attempt to avoid AI Overviews; that removes the site from Google Search.
Bot protection does not block AI search crawlers, which shows up as 403 or 429 responses in logs even when robots.txt allows them.
Logs confirm visits from the allowed crawlers to priority pages.

See AI crawler user agents for the agent list and robots.txt patterns, and AI bot log file analysis for verification.

3. Rendering and JavaScript

Googlebot renders JavaScript. The major AI crawlers from OpenAI, Anthropic, Perplexity and Meta did not in the Vercel and MERJ study.

Main content is present in the raw HTML of priority pages: headings, body text, prices, product details and internal links.
Tabs, accordions and "read more" sections include their content in the HTML rather than loading it on click.
Reviews, ratings and FAQs are not only injected by client-side widgets.
Internal links are real anchor elements with href attributes.
Rendered and raw crawls match closely in word count, headings and links per URL.

See can AI crawlers read JavaScript?.

4. Indexation

Indexed page count is sensible. Compare Search Console's indexed pages with the number of pages that should be indexed.
No important pages are noindexed by meta robots tags or X-Robots-Tag headers.
"Crawled, currently not indexed" and "Discovered, currently not indexed" reasons in Search Console are reviewed for important URLs.
Thin and duplicate pages are consolidated, improved or excluded.
Staging and development sites are not indexable.

5. Canonicalisation and duplication

Every indexable page has a self-referencing canonical unless it deliberately points elsewhere.
Canonicals point to 200, indexable URLs, never to redirects, 404s or noindexed pages.
One version of the site resolves: HTTPS, with or without www, trailing slash policy consistent.
Parameters and session IDs do not create duplicate indexable URLs.
Pagination is handled with crawlable links and sensible canonicals.
Google's chosen canonical matches the declared canonical for important pages, checked with the URL Inspection tool.

6. Site architecture and internal links

Important pages are within three clicks of the homepage.
No orphan pages that are indexable but receive no internal links.
Anchor text is descriptive and consistent for key targets.
Navigation and breadcrumbs reflect the site's structure.
Topic clusters link to and from their hub pages.
No internal links point to redirects or 404s.

7. Redirects and status codes

No redirect chains or loops. Internal links point directly to the final URL.
Permanent moves use 301 or 308, not 302.
Removed pages return 404 or 410, or redirect to a genuinely relevant replacement.
Soft 404s are fixed.
Historic URLs from earlier migrations still redirect correctly. See our website migration SEO checklist.

8. Performance and Core Web Vitals

Field data from the Chrome UX Report or Search Console shows the share of URLs passing LCP, INP and CLS.
Templates, not individual URLs, are identified as the source of failures.
Largest Contentful Paint issues are traced to images, fonts, server response time or render-blocking resources.
Interaction to Next Paint issues are traced to heavy JavaScript or long tasks.
Cumulative Layout Shift issues are traced to images without dimensions, late-loading banners or fonts.
Server response time is reasonable for crawlers as well as users.

9. Mobile

Content and links match between mobile and desktop, since Google indexes mobile-first.
Tap targets, font sizes and viewport settings work on small screens.
Interstitials do not block content on mobile.

10. Structured data

Relevant types are implemented for the client's vertical: Organization, LocalBusiness, Product, Article, Event and others.
Markup validates in the Rich Results Test and Schema Markup Validator.
Markup matches the visible content. Nothing appears only in schema.
Properties are complete and accurate, not minimal placeholders.

Schema helps rich results and entity clarity, but controlled research found it does not increase AI citations on its own. See does schema markup get you cited by AI?.

11. On-page technical elements

Unique, descriptive title tags on every indexable page.
Meta descriptions written for key pages, unique and accurate.
One clear H1 per page, with a logical heading hierarchy below it.
Images have alt text where they carry meaning.
Language and charset are declared.

12. Security

HTTPS everywhere, with no mixed content.
Valid certificate with no upcoming expiry issues.
No hacked content, spam pages or injected links, checked in Search Console's security issues report.

13. International and local

hreflang annotations are reciprocal, use valid language and region codes, and point to canonical URLs.
Each locale has its own crawlable URLs, not content switched by cookies or IP detection alone.
Local business details are consistent across the site and Google Business Profile.

14. Content readiness for AI answers

Not strictly technical, but best checked during the audit:

Priority pages answer their main questions in the first sentences of relevant sections. See answer-first content.
Facts are consistent across pages: prices, durations, locations and service lists.
Pages show an updated date where freshness matters.
The about page states the brand's canonical facts plainly. See entity SEO for LLMs.

How to prioritise the findings

An audit produces far more issues than any team can fix at once. Rank them.

Priority	Typical issues
Critical, fix now	Site or section noindexed or blocked, AI search crawlers blocked, widespread 5xx errors, key content missing from HTML
High	Canonical errors on key templates, redirect chains on important URLs, failing Core Web Vitals on main templates
Medium	Orphan pages, duplicate titles, incomplete structured data
Low	Minor alt text gaps, small numbers of internal links to redirects

Group fixes by template wherever possible. One template fix can resolve thousands of URLs.

Running audits across many sites

For agencies, the value of a checklist multiplies when it is run the same way on every client site, with automated monitoring between full audits. See how to manage SEO for 50+ client websites.

How Rankwyn helps

Rankwyn runs a standard technical and content audit across every connected site, including AI crawler access and raw HTML rendering checks, ranks issues by impact across the whole portfolio, and stages fixes in an approval queue. See the multi-site control plane.

Key takeaways

Audit in order: crawlability, AI crawler access, rendering, indexation, canonicals, architecture, redirects, performance, mobile, structured data, on-page, security and international.
Add two AI-era checks to every audit: AI crawler access rules and content visibility without JavaScript.
Confirm findings with Search Console and server logs, not only crawler output.
Prioritise by impact, and fix templates rather than individual URLs.
Automate monitoring for the issues that cause sudden damage between full audits.

The bottom line

The fundamentals of technical SEO have not changed: pages must be reachable, renderable, indexable and well organised. What has changed is that more systems now read the web, and several of them cannot run JavaScript. Run this checklist on every client site, add the AI access and rendering checks, and prioritise by impact. For the wider AI search strategy, see our AEO and GEO guide.