Technical SEO checklist for a service business website, 2026 edition
Search engines and AI assistants read the same page. The technical work that lets Google crawl, render and trust a service website is the same work that lets ChatGPT, Gemini and Perplexity quote it. This checklist covers crawlability, rendering, speed, structured data, entity consistency, internal links and measurement, with the checks we run on our own site.
ITovio
Editorial team
- A service website needs to be crawlable, rendered as HTML on the server, fast on real phones, and described with structured data; the rest is content.
- Entity consistency (the same name, address, phone and links everywhere) is what lets search engines and AI assistants connect your pages to your company.
- Internal links should follow the buyer's path: problem page to service page to proof page to contact.
- Measure in Search Console and on a fixed set of AI prompts; a change you cannot measure is a guess.
- About one in five people now start a search in an AI assistant; among Gen Z it is roughly one in three (SE Ranking survey, 2026).
- Gartner expects roughly a quarter of organic search traffic to shift toward AI assistants by the end of 2026 (Gartner, via Stackmatix).
- Most AI answers end without a click, which is why being named inside the answer matters more than the referral (Prowlo, 2026).
A service business does not need the technical SEO of a publisher with a million pages. It needs a short list of things done properly and checked on a schedule. This is that list, in the order a crawler meets your site, with the checks we run against itovio.com so you can see what each item looks like in practice.
1. Crawlability: can it be found?
- robots.txt allows the pages you want indexed and blocks only admin, API and thank-you pages. In 2026 it also states your policy for AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) explicitly, so the decision is yours and not the default.
- A sitemap listing every indexable page with a real last-modified date, referenced from robots.txt and submitted in Search Console.
- Canonical tags on every page pointing to the one URL you want indexed, and noindex on pages that should not compete: thank-you pages, internal tools, duplicate category listings.
- Clean status codes: old URLs redirect once (301) to their replacement, not through a chain, and nothing you link to returns 404.
2. Rendering: is the content in the HTML?
Search engines can run JavaScript; AI crawlers mostly do not, and neither waits long. The text, headings and links that matter must be in the server's HTML response. A quick test: fetch the page with curl and search the response for your main heading. If it is not there, it is not reliably indexed. Animations and interactive layers are fine as long as the content underneath is complete without them.
3. Speed: Core Web Vitals on real phones
Google's thresholds are Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds and Cumulative Layout Shift under 0.1, measured in the field. The usual fixes for a service site are small: compress and size images, give every image width and height attributes, preload the one font that paints the headline, serve text compressed with Brotli, and load analytics and chat widgets after the page is interactive and only after consent. Test on a mid-range Android phone over mobile data, because that is where the numbers are worst.
4. Structured data: say what the page is
Structured data is how a page tells machines what it is about without guessing. For a service business the useful set is short:
- Organization on every page, with the same name, logo, phone, email, country and the same list of profiles (LinkedIn, review sites) everywhere.
- Service on each service page, with a description and the provider pointing at the Organization.
- FAQPage wherever you answer real customer questions, because the questions and answers are exactly what AI assistants quote.
- BreadcrumbList on inner pages, Article on blog posts with dates and an author, and WebPage with a speakable selector pointing at the two-sentence summary near the top.
Validate it in the Rich Results Test after every template change. Broken JSON is worse than none.
5. Entity consistency: the same facts everywhere
Search engines and AI assistants build a picture of your company from your site, your profiles and what others write. The picture is only as strong as it is consistent. Use one spelling of the name, one phone number, one email, one description of what you do, and the same links on your site, your Google Business Profile, LinkedIn, review platforms and directories. Publish an llms.txt file that states the facts in plain text for language models, and keep it in step with the site.
6. Internal links: follow the buyer
Link the way a buyer moves: from the problem they feel to the service that fixes it, from the service to the proof (case studies, reviews, a published security posture), and from the proof to contact. Every service page should link to its case studies and its questions; every article should link to the service it supports. Avoid orphan pages, and keep the main navigation short enough that the important pages are one click from everywhere.
7. Security and trust signals
HTTPS with HSTS, the standard security headers, authenticated email (SPF, DKIM, DMARC) and a security.txt file are ranking hygiene and buyer hygiene at the same time. Procurement teams check them; so do crawlers. We publish our own results on the security page for that reason.
8. Measurement
Two instruments. Search Console for impressions, clicks, indexed pages and Core Web Vitals in the field. And a fixed set of buyer prompts run monthly across ChatGPT, Gemini, Perplexity and Google AI Overviews, recording whether your company is named or cited. The second number, share of answer, is the one that tells you whether the structured data, the definitions and the consistency are working where more and more buying decisions now start.
Running the checklist
Most of the list is a one-off fix followed by a quarterly check. The AI Visibility Audit runs it as a fixed two-week scope, with the technical fixes applied, the entity work done and the first share-of-answer baseline measured. The free five-prompt AI visibility check gives you the baseline alone, with screenshots, if you want to see where you stand first.
Questions this article answers
Is technical SEO different for AI assistants?
The foundations are the same: crawlable, server-rendered pages with clear headings and structured data. AI assistants additionally reward definitions, direct answers near the top of a page, cited sources and consistent facts about the company across the web.
Do we need a sitemap if the site is small?
Yes. It costs nothing, it tells crawlers which pages matter and when they changed, and it is the simplest way to find out in Search Console which pages are indexed and which are not.
What Core Web Vitals numbers should we aim for?
Google's thresholds: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1, measured on real visitors' devices, not only in a lab tool.
How often should the checklist be re-run?
Quarterly, and after any redesign, platform change or migration. Most regressions arrive with a change nobody thought was technical, such as a new image carousel or a chat widget.
Share this article
Related Articles
Where the competition stopped, we start.
Security, CRM automation on GoHighLevel, and search visibility. Fixed scopes, written evidence.

