Home / Field guides / GEO & AI Search
Field guide · GEO & AI SearchWhat actually decides whether AI names you.
What the work actually involves, what you can run yourself, and how to tell the difference. Written for owners and marketing leads in the UAE, Saudi Arabia and the wider Gulf.
Written to be used.
A fair number of the brands who read it will decide to run this themselves. That is a good outcome. We would rather be the firm that told you plainly what the job involves than the one that made it sound like alchemy.
Two things have changed at once. Buyers increasingly ask an engine before they ask a person — and the engine names two or three brands, not ten. At the same time, the tools to do this work became widely available, which means the general knowledge is now free and the advantage has moved somewhere else.
You can generate the work. You cannot delegate the outcome.
What you can genuinely do yourself.
- First drafts of anything
- Understanding the concepts
- Keeping your listings accurate
- A first keyword and question map
- Titles and descriptions at small scale
- Editing drafts you already understand
- Which Gulf sources the models use
- A linked graph, not competing blocks
- The order in which citations compound
If you are a single-service business in a market where nobody is competing for the same buyer, you probably do not need a mandate at all. If you tell us that on a call, we will tell you the same thing.
Where the general answer stops being useful.
Not because it is wrong. Because it is general, and this is a contested problem against named competitors who were handed the same advice.
“How do I get cited by AI engines?”
Correct. Also identical to the answer your competitor received this morning.
What the decision requiresWhich third-party sources the models actually draw on for the Gulf — not the ones they draw on for the United States. Which of those you can realistically earn a place in this quarter rather than in theory. Which entity facts must agree across all of them before any of it counts. And the order, because citations compound: sequenced wrongly, two quarters of effort produce nothing measurable.
“Write the schema markup for my business.”
Also correct — and quite possibly harmful.
What the decision requiresWhether you should have two entities at all, or one graph with linked identifiers. What to do about the competing blocks your theme and plugins are already emitting. Whether the licence number, founding date and telephone agree with every third-party listing that mentions you — because an inconsistent entity is worse than no markup. Validation is not correctness.
“Which keywords should I target?”
Volume is the one variable that never pays you.
What the decision requiresWhich of them convert at your margin, not merely at all. Which are held by aggregators and directories you will not outrank and should stop paying to chase. Which are questions people put to an engine rather than type into a search box — different phrasing, different intent, a different page entirely. And which to deliberately abandon.
What a year of doing it alone usually looks like.
- Publish at speed, all competentA busy site, nothing rankingRemoval, harder to authorise than publishing was
- Generate the schema and validate itClean markup describing the wrong companyReconciling licence, address and name to every listing
- Leave redirects to the plugin's defaultsEnquiries had stopped before anyone noticedMapping redirects after the enquiries have gone
- Add Arabic by translationCorrect grammar, phrased nothing like buyers searchHalf the market, invisible
- Choose keywords by volumeTraffic up, enquiries flatAnswering people who were never going to buy
If you run it yourself, run it in this order.
- FirstFix the entityPublishing on an unresolved entity wastes the work
- ThenEstablish the baselineWithout today's answers you cannot prove anything later
- ThenClear the technical groundInvisible when right, and the reason nothing works when wrong
- ThenPublish answerable pages, slowlyPages that deserve to be quoted beat pages that merely exist
- ThenEarn corroborationA credible third-party mention outweighs much of your own publishing
- MonthlyRe-run the promptsNot traffic - the answers are the scoreboard now
No one can guarantee AI answers — we improve the signals that earn them.
Six workstreams, drawn as one machine.
What the PDF adds to this.
- —What sits underneath a short label — one line item opened into the decisions it actually contains.
- —The shape of the job, honestly stated — scope, cadence and what a mandate really asks of you.
- —The whole document in one PDF, formatted to print and to pass around internally.
- —The written argument behind every diagram on this page, in full.
Run it yourself.
Or hand it to someone whose name is on the result.
The tasks became cheap; who answers for the number at the end of the quarter did not change. We will tell you honestly if you are not a fit.
Request a Growth Audit