Why does naming people on my website matter for AI search?

AI engines and Google weigh experience and expertise. A claim like 20 years of experience is stronger when it is tied to a named founder with Person schema and a LinkedIn profile. With no names, the claim has nothing to attach to.

aziz/saif

● Business · SEO · AEO · GEO · Full audit · 19 Oct 2026

A website audit case study: great content, broken plumbing

Q: What is the difference between SEO, AEO and GEO?

SEO is ranking in classic Google and Bing results. AEO (answer engine optimization) is winning featured snippets, People Also Ask boxes and Google AI Overviews. GEO (generative engine optimization) is being quoted and cited by ChatGPT, Perplexity, Gemini and Claude.

Q: What is a redirect loop and how do I find one?

A redirect loop happens when URL A redirects to URL B and URL B redirects back to A. Browsers stop after about 20 hops and show a too many redirects error. Run curl -I on the URL with and without the trailing slash and check that both end in a 200.

Q: Should my sitemap list URLs with or without a trailing slash?

List exactly the URL your canonical tag names. If the canonical has a trailing slash, the sitemap should too. Any other version should 301 redirect to it, not 307.

Q: What is llms.txt and do I need one?

llms.txt is a plain markdown file at the root of a website that summarises who you are, what you do and your key facts, written to be quoted by AI assistants. It costs an hour to write and helps ChatGPT and Perplexity describe you accurately.

Q: How long does it take to fix issues like these?

In this audit the critical fixes (redirect loop, missing pages, share image, dead link) were about half a day of developer work. Schema, FAQ blocks and a founder page took about two more days.

The company is a small software studio in the UAE. Its writing is better than most. It even had an llms.txt file before most of its competitors. But one crawl showed that Google, ChatGPT and real visitors were hitting errors on the pages that mattered most. The company stays anonymous here. Every finding is real.

1. What was audited (and what wasn't)

We crawled 19 pages of the live site on 19 October 2026. For each page we recorded the HTTP status, every redirect hop, the title, description, canonical, robots tag, headings, image alt text and any JSON-LD. We also read robots.txt, sitemap.xml and llms.txt, and fetched the homepage as GPTBot, ClaudeBot, PerplexityBot and Googlebot.

Three things were out of reach because we didn't own the site: Search Console data, live keyword rankings and Core Web Vitals. Nothing in this audit is estimated. Where we had no data, we say so.

Why three letters, not one: SEO gets you into Google's blue links. AEO gets you into the answer box and AI Overviews. GEO gets you quoted by ChatGPT, Perplexity, Gemini and Claude. One site can pass one test and fail the others.

2. Discovery-killers (fix these first)

A discovery-killer is a problem that wastes every other SEO effort. You can write the best service page in Dubai, and if the URL errors, nobody reads it.

Sev	Issue	Evidence	Fix
CRITICAL	About page in a redirect loop	`/about` → 307 → `/about/` → 301 → `/about`. The page sits in the main nav and the sitemap.	Delete one of the two rules. Keep the slash version to match the canonicals.
CRITICAL	Every case study returns 404	All 3 case-study URLs return 404, with or without the slash. They're linked from the homepage, the work page and the sitemap.	Publish them. They're the strongest proof the company has.
CRITICAL	Share image missing	`og:image` and `twitter:image` point at a file that returns 404.	Upload a 1200×630 image at that path.
CRITICAL	Dead "Terms" link	The footer on every page links to a terms page that returns 404.	Publish it or remove the link.
HIGH	Sitemap and canonical disagree	22 inner URLs are listed without a slash. Each 307-redirects (temporary) to the slash version, which is the one the canonical names.	List the canonical URLs. Make the redirect a 301.
HIGH	Two live hosts	The www and non-www versions both return 200. Nothing redirects between them.	One 301 rule at the CDN.
HIGH	Blocked page in the nav	robots.txt disallows a 1,300-word page that the main menu and footer promote.	Unblock it and add it to the sitemap, or stop linking to it.

How the loop happened: the site framework adds a trailing slash. A rule at the CDN removes it. Each rule is correct on its own. Together they bounce the visitor back and forth until the browser stops. Nobody notices until a crawler follows both.

→ Test every nav URL with curl, both with and without the slash

3. SEO findings

Sev	Issue	Evidence
HIGH	No Organization or WebSite schema on the homepage	Zero JSON-LD on the homepage. Only one service page and the blog posts had any.
MEDIUM	Titles too long	75 to 98 characters on the homepage and the four service pages. Google shows about 60.
MEDIUM	Generic titles	Three pages titled just "Page name + brand", 29 characters, no keyword or city.
MEDIUM	Descriptions too long	175 to 200 characters on the service pages. Aim for 140 to 160.
MEDIUM	Duplicate description	The homepage and the work page use the same one.
MEDIUM	Heading errors	The work page has no H1. One blog post has two.
MEDIUM	Thin pages	Four pages under 360 words, one under 160.
MEDIUM	Weak local signals	The homepage shows an email address and nothing else. No phone, no address, no map. The WhatsApp number appears on one page only.
PASS	The basics	HTTPS, one H1 on the main pages, alt text on every image, viewport set, robots.txt points to the sitemap, self-referencing canonicals.

4. AEO findings (answer boxes and AI Overviews)

FAQ markup on 1 page out of 19. A second page had a visible FAQ with no schema. The four service pages had neither.
Headings were slogans. They read well on the page, but nobody searches for a slogan. A heading like "How long does it take to build an MVP in Dubai?" can win an answer box, as long as the first sentence under it gives the number.
No definitional first sentences. Each service page opened with a pitch. Answer engines lift the sentence that starts "X is…".
The process was ready for list markup. Five stages, each with a duration, already sat in the text. Turning that into an ordered list is the cheapest snippet win on the site.
Price was never answered. "How much does it cost" is the most asked buyer question. Even a range would compete for the answer box. Whether to publish one is a business call.

5. GEO findings (ChatGPT, Perplexity, Gemini, Claude)

What worked

A real llms.txt: team size, markets, engagement models, process timings and client results in AED.
All four major AI crawlers got HTTP 200. Nothing was blocked at the CDN.
Concrete, quotable proof numbers: annual savings in AED, a public Google rating, systems merged, products shipped.

What was missing

Not one person named. No founder, no team page, no author bios. The blog author was the company itself. A claim of many years' experience had no person behind it.
One sameAs profile. Only a LinkedIn company page. No Google Business Profile, Clutch, Crunchbase or founder LinkedIn.
The proof didn't resolve. llms.txt described the case studies, and the case-study URLs were 404. An AI engine that checks a claim and finds a dead page trusts the rest less.
Decorative numbers. Dashboard-style widgets showed figures like "runs per day" with no source. They look like data and can't be verified.

The pattern we see most in the UAE: teams invest in copy and design, then never check the site the way a crawler does. The content is fine. The pipes leak.

→ Crawl your own site once a month, not once a year

6. The 10-step fix plan

#	Action	Effort	Helps
1	Break the redirect loop	15 min	All
2	Publish the case studies, or repoint their links until they're live	½ day	SEO · GEO
3	Upload the share image. Fix or remove the terms link	30 min	Social · SEO
4	Sitemap with canonical URLs, 301 not 307, one host	1 hr	SEO
5	Site-wide Organization + WebSite JSON-LD with logo, contact, address, founder and sameAs	1 hr	SEO · GEO
6	Resolve the blocked-but-promoted page	10 min	SEO
7	Founder or team section with Person schema. Real people as blog authors	½ day	GEO
8	5-question FAQ block + FAQPage schema on each service page	1 day	AEO
9	Titles ≤60 characters, descriptions ≤160, fix the H1s	1 hr	SEO
10	Phone, WhatsApp and address in the footer. Claim a Google Business Profile. Add llms-full.txt	2 hr	Local · GEO

7. Check your own site in 10 minutes

Click every item in your main menu. Then type each URL by hand with a trailing slash and again without it. Every one should load.
Paste your homepage URL into WhatsApp, sent to yourself. No image in the preview means your share image is broken.
Open yoursite.com/sitemap.xml and click 5 random links. Any 404 is a page Google is being told to index that doesn't exist.
Open yoursite.com/robots.txt. Check that nothing you promote is under "Disallow".
Open the page source (Ctrl+U), search for "ld+json". None on the homepage means Google has to guess who you are.
Search your page for a real person's name. If there isn't one, AI engines have no expert to credit.
Ask ChatGPT "Who is [your company] in Dubai?" If the answer is vague or wrong, your GEO needs work.

Want this done on your site? I run the same crawl and send a 4-page PDF with every finding and the fix order. WhatsApp +971 55 649 2370 or book a call.

FAQ

What is the difference between SEO, AEO and GEO?

SEO is ranking in Google and Bing's normal results. AEO is winning featured snippets, People Also Ask and Google AI Overviews. GEO is being quoted by ChatGPT, Perplexity, Gemini and Claude.

What is a redirect loop and how do I find one?

URL A redirects to URL B, and B redirects back to A. Browsers give up after about 20 hops. Test each URL with and without the trailing slash. Both should end in a page that loads.

Should my sitemap list URLs with or without a trailing slash?

Exactly the version your canonical tag names. Every other version should 301 to it, not 307.

What is llms.txt and do I need one?

A plain markdown file at your site root that sums up who you are and your key facts, written for AI assistants to quote. It takes about an hour to write.

Why does naming people matter for AI search?

Google and AI engines weigh experience and expertise. An experience claim tied to a named founder, with Person schema and a LinkedIn link, carries weight. With no name, it has nothing to attach to.

How long does it take to fix issues like these?

Here the critical fixes were about half a day of developer work. Schema, FAQ blocks and a founder page took about two more days.