In English, AI names businesses for Saudi buyers about as often as it does in the UAE or the US: 93% of Claude's live-search answers named at least one. In Arabic that fell to 69%, and 31% of Arabic answers named no business at all. The gap holds on like-for-like questions, and it is widest where a district name looks like somewhere else: asked about Al Malqa in Riyadh, the search kept finding Malaga and Malacca.
The answers are built mostly from directories, job ads, classifieds and property portals. Businesses' own websites were only 17% of sources, except in law, where a few firms with Arabic pages for each city and service were named again and again. For a Saudi business the fix is concrete: Arabic pages that answer the exact question, with the district, the city and the price.
Why we ran this
Saudi buyers are moving to AI quickly. Microsoft's AI Diffusion Report puts generative AI use at 31.4% of working-age people in Saudi Arabia in Q2 2026, and the Kingdom climbed from 30th to 25th place in the world that quarter. More and more "who should I use?" questions now end in an AI answer with a short list of names.
On 21 October 2026 we published the same study for the UAE and the US. Saudi Arabia is the third market, with the same assistant, search tool and method, and 350 of its 500 questions were in Arabic.
How we did it
- 500 questions. 10 industries, 50 questions each: 35 in Arabic and 15 in English, written the way Saudi buyers ask, across Riyadh, Jeddah, Dammam, Al Khobar, Makkah and Madinah. The Arabic set also names north Riyadh and three districts (Al Malqa and Al Narjis in Riyadh, Al Rawdah in Jeddah); the English set also names north Riyadh and Al Olaya. Examples: ما أفضل عيادة أسنان في الرياض؟ (What is the best dental clinic in Riyadh?), "Best labour lawyer in Jeddah for expats?", كم سعر فيلر الشفايف في حي النرجس وأين المكان الآمن؟ (How much is lip filler in Al Narjis, and where is safe?).
- One lens: Claude with live web search. Every question was sent once, word for word, to Anthropic's web search tool, as in our UAE and US studies, on 22 October 2026 between 08:13 and 08:39 Riyadh time (05:13 to 05:39 UTC). We saved the answer and every source listed with it: 4,543 sources from 767 websites. All 500 calls worked and every answer ran a search.
- Coding the names. Every specific business the answer named or recommended was listed. Platforms and directories (Bayut, Haraj, Zavis, Doctify and the like), government bodies and generic categories were left out. Named dentists, lawyers and trainers were kept (5.6% of names), and every headline was also computed without them. An answer that named no one, or only pointed to maps, reviews or directories, was coded as hedged.
- Coding the sources. Nine types, the same as the US study. We reused the US site lists unchanged and added the Saudi and Arabic sites, opening each one before deciding what it is.
- Places and a hand check. We read in full every answer that mentioned a place outside Saudi Arabia (148). We also read 20 answers drawn at random against their coded names: of 35 names, none were missed and no platform slipped in; one was a double count (a hotel and the bar inside it), about 3%. The hedged label was right on all 20.
Finding 1: Arabic answers name fewer businesses and hedge far more
Arabic questions got Arabic answers (346 of 350). What was missing was something to name.
| Claude with live web search | English (150) | Arabic (350) |
|---|---|---|
| Answers naming at least one business | 93.3% | 69.4% |
| Businesses per answer, when any | 4.2 | 2.9 |
| Hedged (named no one) | 6.7% | 30.6% |
| Like for like: "best X in [city]", six city names, no price or urgent wording | 93.7% named (143) | 77.8% named (180) |
| Sources that were Arabic pages | 10.8% | 74.6% |
Only the Arabic set had districts and price questions, so the like-for-like row matters. Even without them, Arabic answers named a business 78% of the time against 94% in English, with 3.0 names against 4.2. Arabic was lower in every industry: aesthetics fell from 93% in English to 54%, business setup from 87% to 49%, real estate from 87% to 54%, dental from 100% to 69%.
So the gap is not about Arabic as such. In our UAE study, 88.8% of Arabic answers named a business and 11.6% hedged, against 69.4% and 30.6% here, while English answers were almost the same in both countries (95.2% and 93.3%). Behind Saudi Arabic answers, 74.6% of sources were Arabic pages, against 90.6% in the UAE measured the same way.
Three things in the data explain most of the gap:
- Medical tourism fills Arabic health questions. In Arabic aesthetics answers, 48% of sources were medical tourism platforms listing clinics in Turkey, Egypt, Thailand and elsewhere, in 27 of 35 answers; in Arabic dental answers, 25%. Those platforms publish Arabic pages for every treatment and city, so they match Arabic questions closely.
- Price questions find no prices. 41 questions asked about a price, 38 of them in Arabic. Only 44% named any business and 56% hedged. Several answers offered whitening or filler prices from Dubai or Cairo "as a rough reference" instead.
- District names send the search abroad. Arabic questions naming one of the three districts named a business 60% of the time, against 72% for the six city names.
أفضل محامي طلاق في حي النرجس؟
لم أجد في نتائج البحث أي توصية باسم محامٍ طلاق في حي النرجس. النتائج الأخرى تتناول محامين أمريكيين ومحامين بريطانيين، أو تتحدث عن الطلاق مع شريك نرجسي، وليس لها علاقة بحي النرجس.
Finding 2: Place names that send the search somewhere else
Live search matches the words of the question, and several Saudi place names match other words or places.
| Place in the Arabic question | Answers | Named anyone | Ran into another place | Where the search went |
|---|---|---|---|---|
| Al Malqa (حي الملقا), Riyadh | 40 | 48% | 23 | Malaga and Mallorca in Spain, Malacca in Malaysia, Malaga's football club, clinics in Turkey and Egypt |
| Al Narjis (حي النرجس), Riyadh | 32 | 66% | 17 | "narcissist" divorce pages in Texas, New York and Chicago, El Narges in New Cairo, a Dubai nursery called Narjis |
| Riyadh (الرياض) | 35 | 63% | 6 | dental prices and clinics in Cairo, Dubai and Turkey; for a government-services office, sports news (الرياضة means sport) |
| Al Rawdah (حي الروضة), Jeddah | 35 | 69% | 4 | Al Rawdah areas in Ajman and Sharjah on delivery apps, cafes in Dubai, clinics in Cairo |
| North Riyadh (شمال الرياض) | 29 | 83% | 0 | nowhere: the plainest place name did best |
Al Malqa is the extreme: 53% of those answers named no one, and 28% of their sources were about another country. English had its own traps. Of 19 English questions about Madinah, 11 ran into other places, such as Madinaty in Cairo, Madina Town in Faisalabad, Pakistan, and Medinah, Illinois. Makkah questions found a Texas dental practice called Mecca Dental, Mecca Mall in Amman, Makassar in Indonesia and an agency called Al Makkah Real Estate in Lahore. "Nursery" brought plant nurseries in Jeddah, and "car service" a chauffeur company in Makkah.
Best movers in Madinah?
The search didn't turn up any moving companies based in Madinah, Saudi Arabia. Most results are for other places: Medinah, Illinois: A directory of top-rated movers there, which is a different city despite the similar name.
13 answers were built mostly on another country, 11 of them in Arabic: whitening prices from Dubai and Cairo for Riyadh, a Bangkok clinic for Botox in Jeddah, Malaga's football club and a Sharjah club for "best sports club in Al Malqa". 113 more had results from abroad that the answer noticed and set aside, or one foreign business next to local ones. In all, 26% of Arabic and 23% of English answers were affected, and 20 answers listed a business located outside Saudi Arabia.
What this means for you. Write where you are in full, every time, in both scripts: حي الملقا، الرياض، المملكة العربية السعودية and "Al Malqa, Riyadh, Saudi Arabia". North Riyadh, the plainest place name in our questions, never drifted.
Finding 3: The answers are built from listings, not business websites
Claude listed 9.1 sources per answer:
| Type of source | Share of 4,543 sources |
|---|---|
| Directories, listings, portals, classifieds and job boards | 50.1% |
| Businesses' own websites and other sites | 17.4% |
| News, magazines and Arabic content sites | 15.9% |
| Off-topic hosts (spam-style pages on unrelated sites) | 10.1% |
| Social media and forums | 3.4% |
| Government sites | 1.6% |
| Press releases and syndicated law-firm articles | 0.7% |
| Review platforms | 0.6% |
| Wikipedia and Saudipedia | 0.2% |
86% of answers cited at least one directory or listing. Which kind depends on the industry:
- Health: in dental answers, 78% of sources were directories, led by WhatClinic, Zavis and Bookimed.
- Real estate: property portals (Raghdan, Bayut, Aqar, Wasalt) were 54% of sources.
- Education: school directories (Mdares, Yaschools) were 44% of sources; Mdares appeared in 24 of 50 answers.
- Home services: classified ads from Haraj, OpenSooq, Mstaml and Mourjan were 18% of sources.
- Job ads, everywhere: job boards were 5.7% of all sources and appeared in 99 answers. One Saudi jobs platform, Sabbar, was cited 196 times in 71 answers, for questions such as "best pediatric dentist in Riyadh": the job ad names the clinic and the district.
Businesses' own websites led only in law (57%) and business setup (55%); in every other industry they were 3% to 14% of sources. Only 14% of sources had a "best" or "top" style title, against 38% in the UAE and 26% in the US: Saudi answers lean on raw listings more than on "best of" articles.
Google was advice, not a source. 236 answers, including 52% of Arabic answers, told the buyer to check Google Maps or Google reviews, but only 7 cited a Google business page. Saudi government sites were cited 32 times, fewer than foreign government sites (42), such as Bahrain's ministry of industry and commerce for trademark questions.
Our reading. The search favours pages whose words match the question. In Saudi Arabia those are mostly listings: a job ad for a "pediatric dentist in Riyadh", a Haraj ad for "AC repair north Riyadh". A plain Arabic page on your own site that answers the same question competes with them directly.
Finding 4: In law, Arabic pages for each city win
Law was the exception, and the clearest lesson. 57% of the sources behind law answers were firms' own websites, and the five firms named most often took 52.5% of all mentions. No industry in any of our three studies was this concentrated: the highest top-five share was 15% in the US and 14% in the UAE.
| Law firms named most often in our 50 law answers | Answers |
|---|---|
| Etqan Al Mutamayiza Law Firm | 23 (21 of the 35 Arabic answers) |
| Al Safwa Law Firm | 15 |
| Sadany & Partners | 9 |
| Abdulrahman Al-Mehleky law office | 9 |
| Sahal Law Firm | 7 |
These are counts from our answers, not rankings or recommendations. What the firms share is many Arabic pages, each written for one city and one service. The first firm had 72 different pages cited, for questions about nine places, from "lawyer in Madinah" and "best lawyer in Dammam" to company setup costs; its site appeared in 30 of the 50 law answers and 15 business setup answers. Another firm's pages carried titles such as "best law firm in Al Khobar for 2024", alongside pages for Makkah and Jeddah.
The answers often called these pages the firms' own marketing, and Arabic law answers named only 1.8 firms on average. But they named them: 86% of Arabic law answers named a firm. One firm appeared under 17 different spellings, which shows why a settled Arabic name helps.
The opening for everyone else. In eight industries, businesses' own websites were 3% to 14% of sources. The law firms show what happens when a business publishes an Arabic page for each question its buyers ask, city by city. Outside law and business setup, no business's own site appeared in more than 9 answers; the top law firm's appeared in 45.
Finding 5: One source in ten was a spam-style page
461 sources, 10.1% of the total, were pages about a Saudi business topic on a website that has nothing to do with it, from 150 different sites, in 31% of answers. We call them off-topic hosts, as in the US. Examples, without naming the sites, which are often victims:
- track pages on a music platform titled "best lawyer in Jeddah" and "cleaning company"
- a feedback forum for accounting software with posts such as "chandelier cleaning companies Jeddah"
- a song-lyrics site carrying "best divorce lawyer" articles
- a toothpaste brand's test server with "dentist open now" pages, and free page builders full of "best X in Riyadh" text
Home services were hit hardest: 46% of their sources, and in 22 of 50 home services answers these pages were most of the sources. Aesthetics came next at 20%. Overall, Saudi Arabia (10.1%) sat between the US (15.9%) and the UAE, where we found almost none.
What this means for you. In home services, your competitor for a place in the answer is often a spam page that happens to match the question. A real Arabic page for "AC repair in Al Khobar" with your areas, prices and WhatsApp number competes with it directly. Do not pay for placement on pages like these.
The lists are open, except in law
The answers named 906 different businesses in 1,277 mentions; 80% were named only once (US study: 93%). Outside law and business setup, the five most-named businesses took 10% to 27% of an industry's mentions.
| Industry | Named most often in our answers (number of answers) | Top 5 share of mentions |
|---|---|---|
| Law firms | Etqan Al Mutamayiza (23), Al Safwa (15), Sadany & Partners (9) | 52.5% |
| Business setup and services | Etqan Al Mutamayiza (8), Sadany & Partners (6), Mizan Accounting (6), Quanta (5) | 29.4% |
| Home services | Al Noor company (5), Maskanak (4), Toutach (4) | 26.7% |
| Automotive | Sixt (7), Hertz (6), Enterprise (6) | 22.3% |
| Aesthetics and dermatology | Ram Clinics (4), Ideal Clinics (4), Dalia Clinic (4) | 18.7% |
| Education and childcare | EngVarta (5), British Council (4), British International School Dammam (4) | 13.9% |
| Dental clinics | Kharsa Clinics (6), Maleen Dental Clinic (3), Miral Dental Clinics (3) | 12.2% |
| Fitness and wellness | a physiotherapy centre in Yanbu (4), Taqat Al Liyaqa ladies club (3), Fitness Time (3) | 11.3% |
| Real estate | Irad (3), Retal (3), Asib (3) | 11.0% |
| Restaurants and cafes | Ritz-Carlton hotel restaurants (6), Al Orjouan (5), Arabesque (4) | 10.1% |
These are counts, not rankings or recommendations. In dental answers, 22% of the names were individual dentists.
By industry and city
| Industry | Named anyone, Arabic | Named anyone, English | Names per answer | Hedged | Named once only |
|---|---|---|---|---|---|
| Restaurants and cafes | 91% | 100% | 4.4 | 6% | 84% |
| Law firms | 86% | 100% | 2.7 | 10% | 77% |
| Education and childcare | 80% | 87% | 3.5 | 18% | 83% |
| Fitness and wellness | 71% | 100% | 3.6 | 20% | 87% |
| Dental clinics | 69% | 100% | 3.8 | 22% | 76% |
| Home services | 74% | 80% | 2.0 | 24% | 68% |
| Automotive | 66% | 100% | 3.4 | 24% | 73% |
| Aesthetics and dermatology | 54% | 93% | 2.8 | 34% | 80% |
| Real estate | 54% | 87% | 3.7 | 36% | 89% |
| Business setup and services | 49% | 87% | 3.4 | 40% | 62% |
Each industry had 35 Arabic and 15 English questions, so the English column is directional. By city, Arabic answers named someone 63% of the time in Riyadh (pulled down by Al Malqa and Al Narjis), 65% in Makkah, 71% in Madinah and Al Khobar, 76% in Jeddah and 83% in Dammam. Madinah (37% of answers) and Makkah (29%) ran into other places most.
UAE, US and Saudi Arabia, side by side
Same assistant, search tool and design: 500 questions, 10 industries, two languages. The language mix differs: the UAE study was half Arabic, the US study 80% English with Spanish, this study 70% Arabic.
| Claude with live web search | UAE | US | Saudi Arabia |
|---|---|---|---|
| Answers naming at least one business | 92.0% | 93.8% | 76.6% |
| Same, counting firms only (people removed) | names mostly firms | 92.2% | 74.8% |
| Businesses per answer, when any | 4.3 | 5.2 | 3.3 |
| Hedged | 8.4% | 6.2% | 23.4% |
| English answers naming anyone | 95.2% | 94.5% | 93.3% |
| Arabic answers naming anyone (US: Spanish) | 88.8% | 91.0% | 69.4% |
| Arabic pages behind Arabic answers | 90.6% | not asked | 74.6% |
| Sources per answer | 9.3 | 9.2 | 9.1 |
| Business and other websites (incl. off-topic hosts) | 79.4% | 44.9% | 27.5% |
| of which off-topic hosts | not coded | 15.9% | 10.1% |
| Directories and listings | 9.5% | 37.3% | 50.1% |
| News and media (incl. press releases) | 0.9% | 12.2% | 16.6% |
| Review platforms | 2.9% | 3.2% | 0.6% |
| Social media and forums | 1.3% | 1.6% | 3.4% |
| Wikipedia (Saudi Arabia: with Saudipedia) | 5.8% | 0.1% | 0.2% |
| Government sites | 0.2% | 0.9% | 1.6% |
| Sources with a "best" or "top" style title | 37.7% | 26.3% | 14.3% |
How to read it. The UAE and US figures are as published. Run through the same final lists, business websites come to 72.0% (UAE), 28.6% (US) and 17.4% (Saudi Arabia), and directories to 14.7%, 37.4% and 50.1%, so the order holds. The Arabic-pages row uses one measure for both countries; the UAE page's 90.2% used a narrower rule.
In all three, English answers name someone almost every time. What differs is the material: businesses' own sites in the UAE, directories and city media in the US, and listings of every kind in Saudi Arabia, where Arabic answers have the least to work with.
What a Saudi business should do with this
- Write the Arabic page your buyer's question needs. One page per service and city, in Arabic written for Saudi readers, not translated: the service, the district, prices in SAR, hours, insurers or licences, and a WhatsApp number.
- Spell out where you are, every time. District, city and country, in Arabic and English, on every page and profile.
- Put prices on the page. 56% of price questions got no business named. A "from" price in SAR gives the answer something local to quote.
- Fix the listings answers read for your category. Clinics: Zavis, WhatClinic, Doctify and Vezeeta. Property: Bayut, Aqar, Raghdan and Wasalt. Schools: Mdares and Yaschools. Car services: Syarah, Motory and YallaMotor. Home services: Haraj and OpenSooq. Spas: Fresha. Same Arabic name, district, phone and prices everywhere.
- Settle one Arabic spelling of your name and use it next to the Latin name on your site, your Google Business Profile and every listing.
- Name your doctors and lawyers and say on each profile where they work.
- Keep Google Maps right. Half the Arabic answers sent the buyer there to check.
- Check in Arabic, by district, on more than one assistant. Ask the same 10 buyer questions on ChatGPT, Gemini and others each month, with your district in the question, and keep dated screenshots.
If you want help
This is the work we do for Saudi businesses, remotely from Dubai, in Arabic and English, with traditional SEO included and ChatGPT and Gemini (Google AI) tracked on every plan. It starts with a free AI check: 10 real buyer questions, with the competitors AI names instead of you. The AI Visibility Audit (30 questions per language, 4 AI tools, 3 competitors, screenshots and a 90-day plan) is AED 1,950 (about SAR 1,990), credited in full to month one if you sign within 30 days. Monthly plans are AED 4,500 (about SAR 4,600) for Foundation, AED 8,500 (about SAR 8,680) for Growth and AED 16,000 (about SAR 16,340) for Authority, month to month, with no lock-in and no setup fee. Prices are fixed in AED; SAR figures are approximate, at the UAE Central Bank rate of 21 October 2026 (1 SAR = AED 0.97902). Details are on our Saudi Arabia page.
In the first 30 days of any plan, if AI does not name you for at least 3 commercial questions on the list you approved, or you are unhappy for any other reason, we refund 100% of what you paid, in cash, within 7 working days. The rules are on our guarantee page. What we don't promise: a top spot on any assistant, guaranteed rankings, paid places on "best of" lists, or a Saudi office. We do not have one.
Limits of this study
- One assistant, one morning (22 October 2026) and one run per question. Answers change from day to day and between assistants. ChatGPT and Gemini carry most AI chatbot traffic in Saudi Arabia, and Claude's search uses its own index, so this is not what every buyer sees.
- English was 150 questions, 15 per industry and 18 or 19 per place, so English figures by industry and city are directional.
- Source types were coded by us. We opened the Saudi and Arabic sites and checked every site with two or more citations by hand, but some single-citation sites may be in the wrong group. The off-topic and "another place" shares are floors.
- Names were matched with simple rules and a hand list of Arabic and Latin spellings, so "named once" may be slightly high. The text checks behind "marketing" warnings miss some phrasings.
- We wrote the questions. We did not check whether every named business exists, is licensed or is still open, and being named is not an endorsement by us or by the assistant.
Downloads
Both files are free to use with credit to Adjoltz and a link to this page.
- The 500 questions (JSON lines): id, industry, language, city or district, and the exact question text.
- One row per question (CSV): place in Arabic and English, question, businesses named (with and without people), hedged, the first business named, Arabic sources, sources about another country, drift, and sources of each of the nine types.
The same method in other markets: the UAE study and the US study. This study in Arabic: دراسة الظهور في الذكاء الاصطناعي في السعودية 2026.