A technical SEO checklist for a UAE website in 2026 should cover ten things: crawling and indexing basics, AI crawler access, mobile Core Web Vitals, server speed, JavaScript rendering, bilingual (Arabic and English) setup, canonical and duplicate URLs, redirects and broken links, structured data, and internal linking. The first two decide whether Google and AI tools can reach your pages at all, so check them first.
Technical SEO is unglamorous, and that’s the point. A site can have excellent content and still lose to a competitor because a robots.txt line, a firewall rule, or a missing language tag stops it from being read properly. We can’t tell you what percentage of UAE sites have these problems, because we haven’t found a published audit that measures it, and we’d rather not invent one. What we can give you is the checklist, in the order we’d work through it.
Quick Answer
Start with whether pages can be crawled and indexed (robots.txt, noindex tags, XML sitemap, Search Console’s page indexing report), then confirm AI crawlers aren’t blocked by your robots.txt or firewall. After that, fix mobile Core Web Vitals (LCP 2.5 seconds or less, INP 200 milliseconds or less, CLS 0.1 or less), make sure key content is visible without JavaScript, set up Arabic and English versions with reciprocal hreflang, and clean up canonicals, redirects, structured data, and internal links.
The 10-Point Technical SEO Checklist at a Glance
| # | Fix | Quick check | Free tool |
| 1 | Crawling and indexing | Are important pages indexed? Any stray noindex or robots.txt blocks? | Search Console |
| 2 | AI crawler access | Do robots.txt and your firewall allow the AI search bots you want? | robots.txt, CDN logs |
| 3 | Core Web Vitals | Do mobile pages pass LCP, INP, and CLS at the 75th percentile? | PageSpeed Insights, Search Console |
| 4 | Server response | Is time to first byte slow for visitors in the Gulf? | PageSpeed Insights |
| 5 | JavaScript rendering | Is key content in the raw HTML? | URL Inspection, view-source |
| 6 | Bilingual setup | Do English and Arabic pages point at each other with hreflang? | Crawler, Search Console |
| 7 | Canonicals and duplicates | Does each page resolve to one preferred URL? | Crawler |
| 8 | Redirects and broken links | Any chains, loops, 404s, or soft 404s? | Crawler, Search Console |
| 9 | Structured data | Does markup validate and match the visible page? | Rich Results Test |
| 10 | Internal linking | Are valuable pages linked from somewhere sensible? | Crawler |
1. Can Google Crawl and Index Your Important Pages?
Open the page indexing report in Google Search Console and confirm that the pages you want to rank are indexed, then look for the usual culprits: a noindex tag left on after a redesign, a robots.txt rule blocking a folder, or an XML sitemap listing URLs that redirect or error.
A single syntax error in robots.txt can block your whole site from Googlebot, which removes you from traditional search and from AI Overviews at the same time. Check that your sitemap lists only canonical, indexable, 200-status URLs, and that it’s submitted in Search Console. While you’re there, confirm the whole site runs on HTTPS with no mixed-content warnings, and that cookie banners or WhatsApp pop-ups aren’t covering the page on mobile.
2. Are AI Crawlers Blocked by Your robots.txt or Firewall?
Allow the retrieval crawlers that feed AI search answers, decide separately about training crawlers, and check your CDN or firewall rules, because a “block unknown bots” setting written before AI search existed can quietly block them all.
This is the check most older checklists don’t have, and it’s easy to get wrong because the bot names look similar:
- OpenAI: OAI-SearchBot powers ChatGPT’s search results and citations. GPTBot is the training crawler. Blocking GPTBot doesn’t remove you from ChatGPT search; blocking OAI-SearchBot does.
- Google: there’s no separate retrieval crawler for AI Overviews or AI Mode. They use the standard Googlebot index. Google-Extended only controls whether Google uses your content for Gemini training and grounding, and it doesn’t affect Search, AI Overviews, or AI Mode.
- Perplexity and Anthropic: PerplexityBot and Claude-SearchBot are retrieval crawlers, with separate training and user-triggered agents. Bot names change, so check each vendor’s current documentation before editing.
Here’s an example of a setup for a business that wants to be cited but opts out of OpenAI training. Treat it as an illustration, not a recommendation, because whether to allow training is your decision:
User-agent: OAI-SearchBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: GPTBot
Disallow: /
Two caveats keep this honest. First, robots.txt is a voluntary signal, not a lock, and some user-triggered fetches may not follow it. Second, if your site sits behind Cloudflare or a similar service, check the firewall and bot-management rules as well as robots.txt. To see whether any of this translates into visibility, use Search Console’s Generative AI performance report for Google’s AI features, and see our guide to the best AI search rank tracking and visibility tools for monitoring ChatGPT, Perplexity, and others. Google’s own guide to generative AI search also says you don’t need an llms.txt file or special markup to appear in its AI features, so don’t spend budget there for Google’s sake.
3. Do Your Mobile Core Web Vitals Pass?
Core Web Vitals are three real-user metrics measured at the 75th percentile of visits: Largest Contentful Paint (LCP) at 2.5 seconds or less, Interaction to Next Paint (INP) at 200 milliseconds or less, and Cumulative Layout Shift (CLS) at 0.1 or less. Check the mobile report first, because Google indexes the mobile version of your site.
INP replaced First Input Delay in March 2024, so any checklist still talking about FID is out of date. You may also have seen claims that Google tightened the LCP threshold to 2 seconds; as of this writing, Google’s own documentation still says 2.5. One 2026 analysis of Chrome UX Report data put the share of sites passing all three metrics at around 55% globally, so passing is common but far from universal.
Be realistic about weight, too. Core Web Vitals are a page-experience signal, not a trump card; they matter most when two pages are otherwise similar, and they matter for conversions regardless.
For Gulf sites, two common culprits are worth checking:
- Arabic web fonts: Arabic font files can be large. Subset them, serve them in WOFF2, and use font-display: swap so text appears before the font finishes loading, which helps both LCP and layout shift.
- Hero images and sliders: oversized banner images and auto-rotating sliders are the usual cause of slow LCP on corporate sites.
4. Is Your Server Fast for Visitors in the Gulf?
Slow server response caps everything else. If time to first byte is high for visitors in the UAE and Saudi Arabia, no amount of image compression will fully fix LCP, so check hosting location, caching, and whether a CDN with a Middle East presence is serving your pages.
Run PageSpeed Insights on your key pages and look at server response time. If you’re hosted far from your audience with no CDN or page caching, that’s the first fix. A well-configured CDN usually matters more than the physical location of the origin server.
5. Is Your Key Content Visible Without JavaScript?
Confirm that your headings, body text, and links appear in the raw HTML, not only after JavaScript runs. Googlebot renders JavaScript, but several AI crawlers have been reported not to, so content that only appears after scripts execute may be invisible to them.
Use “view-source” on a key page, or the URL Inspection tool in Search Console, and check that your main content is there. Single-page apps and heavy front-end frameworks are the usual offenders. Where it matters, server-side rendering or pre-rendering is the fix.
6. Is Your Arabic and English Setup Correct?
Give each language its own crawlable URL, add reciprocal hreflang tags (including a self-reference and an x-default), and declare language and direction in the HTML, because one-way or missing hreflang is the most common bilingual error.
Also check that your Arabic pages have translated titles, meta descriptions, and alt text, not just body copy, and that a language toggle isn’t swapping text on a single URL. We cover this in depth, including a working hreflang example, in our guide to Arabic and bilingual generative AI optimization for Saudi and UAE enterprise search.
7. Does Every Page Resolve to One Preferred URL?
Pick one version of each URL and make sure everything else points to it with a canonical tag or redirect. Duplicate versions (http and https, www and non-www, trailing slash and none, tracking parameters) split signals and waste crawl budget.
Two UAE-specific wrinkles: if you use Arabic characters in URLs, be consistent about it, because they appear percent-encoded and inconsistent handling creates duplicates; and if you run both a .ae domain and a global domain, make sure they’re not serving near-identical content without hreflang or canonical tags that explain the relationship.
8. Are There Redirect Chains, Broken Links, or Soft 404s?
Crawl the site and fix redirect chains and loops, internal links pointing at 404 pages, and soft 404s (pages that say “not found” but return a 200 status). Redesigns and migrations are the main cause. Old URLs get dropped, redirects get skipped, and rankings quietly slide. Map every old URL to its closest new equivalent with a single 301 redirect, then re-crawl to confirm.
9. Does Your Structured Data Validate and Match the Page?
Validate your markup with Google’s Rich Results Test and the Schema Markup Validator, and make sure every marked-up fact is visible on the page. Markup that doesn’t match the visible content is a quality problem, not a harmless leftover.
Check Organization on the homepage, LocalBusiness on genuine location pages, and FAQPage or Article where relevant. Remember that schema helps machines read a page; it doesn’t directly boost rankings. We explain what it does and doesn’t do, and how to keep one entity across English and Arabic, in our Entity SEO & Schema Markup Strategy for Middle East Enterprises.
10. Are Your Valuable Pages Linked From Somewhere Sensible?
Every page you want to rank should be reachable within a few clicks from the homepage and linked from related pages with descriptive anchor text. Orphan pages, which have no internal links pointing to them, are hard for crawlers to find and easy to overlook.
A crawler such as Screaming Frog will list pages with zero internal links. Link services pages to the case studies and guides that support them, and link guides back to the service pages. Avoid stuffing links into footers and menus to compensate; relevant links in the body of the content do more.
How Should You Prioritize These Fixes?
Fix in this order: anything that blocks crawling or indexing (1, 2, 8), then anything that harms the whole site’s experience (3, 4, 5), then structure and clarity (6, 7, 9, 10). A blocked page can’t benefit from any other improvement, so a robots.txt or firewall problem always goes first. Re-test after each change rather than batching ten fixes into one release, so you know which one worked.
Bottom Line
Technical SEO is the layer that decides whether your content gets a fair chance at all. In 2026 that includes a question most checklists still skip: whether AI search crawlers can reach you. Work through the ten checks in order, fix what blocks access first, and re-test as you go.
Need a Second Pair of Eyes on Your Site?
Glomm runs technical SEO audits for UAE businesses, covering crawling, Core Web Vitals, bilingual setup, structured data, and AI crawler access. Learn more about our SEO services or contact Glomm.
At minimum, after every redesign, migration, or CMS change, and quarterly otherwise. Set up Search Console alerts so indexing and Core Web Vitals problems reach you sooner than your next audit.
Yes, as a page-experience signal, but a modest one. Content relevance matters more. They also affect conversions directly, which is reason enough to fix them.
It’s a business decision. Blocking retrieval crawlers like OAI-SearchBot removes you from that tool’s answers, while blocking training crawlers like GPTBot opts out of model training without affecting ChatGPT search. Whatever you choose, make it deliberate rather than the accidental result of an old firewall rule.
No. A country-code domain is a geo signal, but a global domain with clear language and location signals can rank too. Consistency matters more than the domain choice.
Many of these checks are free with Search Console and PageSpeed Insights. Server, JavaScript rendering, and bilingual setup issues usually need a developer.
