Who should analyse your site in AI answers in Saudi Arabia?

Three kinds of provider, each with what it offers and what it does not:

• An analysis platform: it asks the engines your customers' questions on a regular rhythm and says, in each answer, whether your site was cited, who was cited instead of you, why, and what to change first. The work itself stays with you or your team.
• An agency: it makes the edits: writing pages, technical work and getting your name mentioned in other sources. It may measure with a tool, or it may not.
• An independent consultant: reviews your site and sets out a plan, and sometimes carries out part of it. Their value is their knowledge of your sector and market; their limit is their time.

The SEO company you already work with may be enough, as long as it measures the effect of its work in AI answers on your customers' Arabic questions. Most GEO work is editing page text on a technical foundation SEO people know well, and Google says optimising for its generative AI features is still SEO.

What GEO adds is measurement by question rather than by keyword, and knowing who is cited instead of you on each engine, and why. So whoever searches for «شركة سيو في الرياض» ("an SEO company in Riyadh") or «خبير سيو في السعودية» ("an SEO expert in Saudi Arabia") to lead ChatGPT's and Google's answers should ask one question before signing: how will you measure whether the engine cited my site, and on which questions? The difference between the two kinds of work is set out in SEO vs GEO vs AEO.

Ask for three things before any contract: a measurement of your site on your customers' questions in their dialect and city; every answer's excerpt as evidence under each verdict, with the links the engine gave; and a measurement after the work on the same questions, with a check that the edit was published before any effect is credited to it.

The criteria for choosing the measurement software itself, from language and market to a fixed, published price, are in the best tool to analyse where your site stands in AI answers, and what a report worth asking for looks like is in a real report for a clinic group in Riyadh and Jeddah.

Put these seven questions to every company, and ask for the answers in writing:

• Will you measure my site before you start, on my customers' questions and in their dialect, and give me the result with excerpts of the answers?
• Who writes the pages in Arabic, and do they know my market and sector, or do they translate?
• What will you change, on which page and why, before you change it?
• How will you prove the effect of your work: do you check the edit is published, and measure the same question afterwards?
• How many pages a month does the offer cover, and who reviews them before me?
• What stays mine if I stop: the pages written, and the measurement history?
• Is the price fixed and in writing, or are fees added for each question or market?

A perfume shop in Riyadh receives two offers. The first promises it "first place in ChatGPT within a month" and sends a monthly report of screenshots of answers. The second starts by measuring twenty questions the shop's customers write, in their dialect and city, then says a competitor leads seven of them, explains what that competitor's page has that the shop's lacks, and proposes editing three pages first.

The first offer sells a promise it cannot keep, with a report nobody can check. The second sells work that can be measured: the shop knows where it stands before it pays, and will know after the edits whether anything changed. Even if the second costs more, it is the offer that can prove what it added.

And if the shop wants to check the second offer for itself, it measures the same questions with an independent tool before and after the edits, as in how to know if ChatGPT and Gemini cite your site.

Trust here rests on evidence you can open yourself, not on client counts or logos. Four steps before any contract:

• Ask for a past measurement with its dates: a real Arabic question, and the text of the answer before and after their work, with both dates.
• Ask them to measure your site before the contract: on ten of your customers' questions, so you see how they work before you pay.
• Ask about your market: how your customer asks in Riyadh, how that differs from Jeddah or Dubai, and who leads your sector today.
• Ask their clients: ask for one or two clients in Saudi Arabia you can call.

Nobody can guarantee you a place in a ChatGPT, Gemini or Google answer. The engine picks its sources afresh every time, and nobody sells a place among the sources it cites. Whoever guarantees you "first place in ChatGPT" is promising something they do not control. And Google says its generative AI features need no special optimisation and no special files, so there is no "secret technique" anyone can sell for them.

What can be guaranteed is the work itself, so ask for it in writing:

• The number of pages written or edited, and when they are delivered.
• A measurement before and after the work, on agreed questions, with excerpts of the answers.
• A regular report that says what changed, what of it held, and what did not.
• A check that every edit is published before any effect is credited to it.

Tasadr does not do the work of an agency or a consultant, but it does not stop at measuring. We ask the engines your customers' questions in their words and dialect, record one of four outcomes for your site in every answer, and show you who leads instead of you and why, and what to change first.

Every competitor analysis comes with on-page fixes ready to publish: a paragraph that answers the question, an FAQ and structured data (schema), from your page and your business profile only.

The work itself stays with you or whoever you choose: you send the task to a developer, an SEO lead or a content writer in one step, and we check the edit is live on the page before we measure its impact. If you work with an agency, Tasadr is the independent measurement that says whether its work paid off. If you are an agency, how to track your clients' sites is in Tasadr for agencies, and how to build and price the service without promises you cannot keep is in how can your agency sell GEO without empty promises?

You start with a 7-day free trial, with no credit card, for 1 site and 10 questions on ChatGPT and Gemini. All the plans are on Tasadr's pricing.

Six mistakes that cost more than any offer:

• Choosing on price alone, before you know which part of the work it buys.
• Accepting the agency's report about itself as the only measure of its success.
• Starting with no measurement before the work, so nobody can prove later what it added.
• Believing a guarantee of a place in the answer.
• Choosing someone who does not write in Arabic, for a market that asks in Arabic.
• Signing a contract that does not say what stays yours if you stop.

No tool, agency or consultant can guarantee you a citation, or say when the engine will rebuild its answers. What you can ask of anyone is an honest measurement of where you stand today, evidence for every verdict, what to change first, and then a measurement that says whether anything changed. How to read that measurement is in how to measure the success of optimising for AI answers.