Search is changing, but the basics of getting a website noticed have not disappeared. Google still needs to crawl and index pages, visitors still expect useful information, and a slow or confusing website can still drive people away. What has changed is where people find answers.
Today most people ask ChatGPT, Perplexity and Gemini what to buy and who to hire, and those tools only name a handful of brands per answer. If yours isn’t one of them, a page-one ranking won’t feel like much of a win.
So this SEO audit checklist 2026 covers three things: your technical setup, your content, and what I’ll call GEO readiness (generative engine optimization, which really just means being easy for AI tools to find and quote).
This guide takes you through the checks that matter, explains what to look for and helps you decide which problems deserve attention first.
What Actually Changed?
Audits used to ask one question. Can this page rank? These days there’s a longer one: can an AI system fetch your page, make sense of it, trust it, and quote it without getting it wrong?
Any of those can break. Block the crawler and nothing gets fetched. Bury the answer under 600 words of warm-up and nothing gets quoted. Say you’re based in one city on your website and another on LinkedIn, and trust takes a knock.
The reassuring part is the overlap with normal SEO. Google’s own advice still leans on the established basics, plus making sure crawling is allowed. You’re stacking a layer on top, not rebuilding the house. Nobody outside the labs fully knows how these tools pick their sources, mind you, so take anyone’s “secret formula” with a pinch of salt.
Step 1: Start With a Technical SEO Audit
If a bot can’t reach your page, it can’t rank it or quote it. A technical SEO audit is the unglamorous part, but everything else sits on top of it. Search Console, a crawler like Screaming Frog and PageSpeed Insights will get you most of the way there.
Begin with access. Open robots.txt, then go and check your CDN and firewall rules too. A WAF setting can quietly block bots while robots.txt looks perfectly innocent, and I’ve seen that catch people out after a hosting move.
Indexing is next. Your important pages should return a 200 status, be indexable, and appear in your XML sitemap. Open the “Not indexed” report in Search Console. A service page marked “Crawled – currently not indexed” deserves a proper look. A few old tag pages sitting there probably don’t.
Then there’s rendering, which people skip. Open the web page and right click, View Source, and look for the prices or details of your products. Some crawlers may never see the text if it’s not included in the initial HTML, but is added by JavaScript. It is safer to go the server side rendering or pre-rendering route.
The rest is housekeeping, and nobody enjoys it. Do it anyway.
– Load your key pages on an actual phone, on mobile data rather than wifi, not in the desktop preview. An LCP less than 2.5 seconds, an INP less than 200 milliseconds and a CLS less than 0.1. Open the Core Web Vitals report in Search Console, which displays what actual users experienced, in addition to lab tools.
– Afterwards, crawl the site. Screaming Frog does. What you’re searching for are orphan pages, duplicate pages, thin pages, broken links, A-to-B-to-C links. All important should be within 3 clicks from the home page and if further away, should link to it closer to the top.
– Schema comes after that, and only where it fits the page: Organization, Article, Product, FAQ, HowTo, Breadcrumb, LocalBusiness. Don’t mark up anything a visitor can’t see on the page. Run it through Google’s Rich Results Test and fix whatever it flags.
– Last, the sitemap. Keep it current, and submit it in both Google Search Console and Bing Webmaster Tools.
One tip: sort everything you find into three piles. “Blocks traffic,” “hurts traffic,” “nice to have.” Clear the first pile before you touch the third. Audit tools love giving you a score out of 100, but a 92 doesn’t mean anybody will cite you.
Step 2: Go Through Your Content (Be a Bit Ruthless)
Pages pile up. After five years a typical site has a few brilliant posts, a lot of “fine” ones, and some nobody remembers writing. Dump every URL into a sheet with its traffic, rankings and last update date. Then tag each one: keep, update, merge or delete. Brutal? Slightly. Fast? Very.
For the keepers, look at the opening first. Does it directly address the question? Provide a straightforward one or two-sentence response within the first 40-60 words of every key paragraph and then expand the detail later. Headings help. “How much does a website redesign cost?” does more work than “Pricing Overview,” because it matches the way people type, and the way chatbots get asked.
Each section should survive being copied out on its own. Picture an AI tool lifting one block. Would it still be relevant? Maintain an order for your heading: one H1, one or more H2s for major themes, one or more H3s beneath for further questions. Comparison tables, FAQs and step lists belong in real HTML text. A pricing table saved as a screenshot is invisible to a machine.
Specifics beat adjectives every time. “Cuts load time by 40%” is something a tool can quote. “Significantly faster” is just noise.
Then ask the awkward question: why would anyone cite this page? If it says what the top ten results already say, it’s replaceable. Imagine a boiler engineer who publishes what repairs really cost across 40 jobs last winter, photos included. That’s hard to copy. Google points to this kind of non-commodity content when it talks about its generative features, so it’s not just my opinion. First-hand experience, your own numbers, a sharper framework or an honest comparison will all do the job.
Freshness matters too. Anything involving prices, laws, specs or statistics goes stale quickly, so review those pages on a schedule. Add author names, credentials, sources, and both publish and update dates. Be mindful of cannibalization (where two of your pages compete for the same keyword). Combine them together and then 301 the weaker one.
Practical tip: begin with your top 10 most important pages. Update the figures in the opening and introduce one actual customer or from your team.
Step 3: Run an AI Search Readiness Audit
Here’s the new bit. An AI search readiness audit asks whether ChatGPT, Gemini, Perplexity, Claude and Google’s AI Overviews can find you, describe you correctly, and point people your way. Nobody has the full formula yet, so read what follows as sensible bets, not guarantees.
Decide your crawler policy. Search robots.txt and your CDN rules for GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot and Google-Extended. Blocking them can be a perfectly fair business decision. Blocking them by accident isn’t, and blocking them “just in case” is the worst version. Choose a policy, and record it where you will be able to see it later.
Ensure that your brand is easily recognisable. Uniformly name, describe and use your services throughout your website, Google Business Profile, LinkedIn, directories and social profiles, with your name as well as location and contact information. .
Add Organization schema, with sameAs links pointing to your official profiles. Clear About, Services, Pricing, FAQ and Contact pages help a lot here, because they spell out what you do and who for.
Earn mentions off your own site. Reviews, industry publications, podcasts, directories, expert roundups. AI tools seem to draw on a wider web of trust signals than backlinks alone, though how much weight each one carries is anyone’s guess.
Cover the whole buyer journey. People search for alternatives, pricing and fixes, not only the big category term. Build a few “best answer” assets such as comparison pages, calculators, templates or a properly detailed FAQ.
Check how AI describes you. If it gets your services, pricing or location wrong, publish a clearer page and correct the inaccurate listings you can find.
A quick word on llms.txt. Google has said it neither helps nor harms its own visibility, so file it under “optional.”
Step 4: Test Your AI Visibility
You can’t improve what you don’t measure, and measuring is the second half of any AI search readiness audit. Simply keep it to a monthly scorecard.
Begin with 5–10 questions that resemble actual customer questions. When this becomes comfortable, expand to 20 or 40, adding a variety of discovery, comparison, pricing and “alternatives to” questions. Feed them into ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews. Write down for each answer if you were mentioned, the URL that was cited, if you were described accurately, and which competitor did appear. That last one stings a little, but it’s the most useful column.
A few numbers are worth tracking over time:
– AI citation share. How often your brand or domain appears across your priority prompts.
– Branded accuracy. Whether the AI gets your business right.
– Cited-page performance. Which pages get referenced, so you can make more like them.
– AI referral traffic. Some visits arrive with no normal referrer, so set up custom channel grouping and UTM tags where you can.
– Conversion quality. Do those visitors book, buy or fill in a form?
– Content freshness. How long ago was the last time you checked your key pages?
A 30 Days SEO Audit Checklist 2026 Plan
An audit sitting in a spreadsheet does nothing. So here’s a month, roughly.
Week 1. Do your technical SEO audit, resolve indexation and crawler access issues, and select the 10 most important pages for a commercial perspective. .
Week 2. Rewrite the openings on those ten pages. Add direct answers, FAQs, comparison tables and sensible internal links.
Week 3. Add or correct schema, tighten author and organization details, and update the facts that have gone out of date.
Week 4. Conduct first AI search readiness audit prompts, record citations, identify gaps relative to competitors and create a content roadmap. Initial results will be “crude. That’s normal.
Mistakes Worth Avoiding
Treating the audit as a one-off is the big one. For full pass, do this quarterly, for light check do this monthly as one redesign can undo a year of careful work.
Trying to fix every page is another. Start with the pages closest to revenue and let the rest wait.
Then there’s writing for robots. Write for people first. The AI tools were trained based on content that has been written well and helpful by humans, so why would you be trying to chase after them with stiff content?
Finally, if you are having the same problem over and over, consider reviewing your publishing practices.
Wrapping up
SEO is still here. What’s changed is the job. Ranking a page used to be the whole game, but now plenty of people ask ChatGPT or Gemini first, and what they get back depends on whether those tools trust you enough to name you.
That trust comes from fairly ordinary things. A site that works. Pages that answer what your customers actually ask. Proof nobody else can copy, like your own data, your own photos, a result from a real job. Once that’s in place, go and ask the AI tools about your business and see whether your name turns up. Odds are it won’t, the first time you check.
Don’t try to fix everything this week. Pick one thing from the checklist, do it today, and take the next one tomorrow.
+971 582 871635
+971 582 871635
