.webp)
The most credible thing an agency can do is demonstrate its own methodology on its own brand. Anyone can publish a framework. Far fewer can show you the actual SERP, the actual AI Overview, and the exact steps that produced both. This is that piece.
Over the past few months, we set a specific target: rank SellThru's LLM SEO page in the top three positions on Google for "seo llm agency dubai" and "best seo llm agency dubai" — and get cited by Gemini's AI Overview for the same queries. Both happened. Here is the breakdown.
"Ranking in AI is not a separate discipline from ranking on Google. It is the same discipline, measured through a new lens — and the lens is now visible."
The Six Things We Did
There was no single lever. The result came from six elements working together, each reinforcing the others.
Built a dedicated, keyword-specific landing page
We built a standalone page — sellthruagency.com/seo-agency/seo-llm-marketing-agency — specifically targeting "LLM SEO agency Dubai" and related variants. A dedicated page can own a narrow topic completely in a way a general services page never can.
/seo-agency/seo-llm-marketing-agency nests under the parent SEO category, inheriting existing domain authority while building its own topical relevance for LLM-specific queries.Wrote content structured for how AI reads, not just how humans browse
LLMs extract meaning differently from humans. They look for definitional clarity, entity relationships, and citable claims. The page was written with this in mind: clear definitions, explicit statements of what SellThru does and for which markets, and structured claims throughout. Every paragraph was written to be extractable as a standalone answer.
Implemented FAQ schema targeting AI Overview trigger queries
FAQ schema is one of the most direct bridges between page content and AI Overview inclusion. We identified the specific questions users ask around LLM SEO in the Dubai and UAE context and structured the FAQ section to answer each with a complete, citable response. Google's AI Overview pulls directly from well-matched FAQ content.
Applied full structured data — Organisation, Service, and LocalBusiness schema
Beyond FAQ schema, implementation included Organisation schema with explicit service descriptions, LocalBusiness schema with Dubai location data, and Service schema for LLM SEO as a named offering. Structured data tells AI models not just what a page says, but what a business is — with machine-readable certainty that body copy alone cannot provide.
Published an llms.txt file
The llms.txt file — modelled on robots.txt — tells AI crawlers which content on a site is authoritative, how to interpret the site's structure, and which pages represent the primary service offering. We implemented this when adoption among Dubai agencies was effectively zero.
Built targeted backlinks from contextually relevant sources
Backlinks remain a trust signal for both Google's algorithm and, indirectly, for LLMs that weight content from cited and linked domains more heavily. We pursued a small number of high-relevance placements: industry directories, agency listing sites covering the MENA digital marketing space, and contextual mentions in content covering AI and SEO in the Gulf.
What's Coming Next: The Infrastructure Behind It
The six tactics above are executable by any disciplined SEO team. What we're building at SellThru goes further — to run this playbook for clients at scale, with consistent methodology, measurable outputs, and the ability to track LLM visibility over time.
The infrastructure we've been building connects live data sources — DataForSEO, Google Search Console, GA4 — to an AI analysis layer powered by Claude (Anthropic's model), running through our own hosted MCP server. The output is a family of structured audit deliverables, with the GEO/LLM visibility audit using active querying against ChatGPT and Gemini APIs. A full breakdown is coming in a dedicated piece.
The GEO/LLM Audit output (crimson) is the end-product most directly connected to the ranking work described in this article. Active querying via ChatGPT and Gemini APIs is the only valid methodology for MENA markets — passive monitoring endpoints return empty for GCC by design. A full technical breakdown of this infrastructure will follow in a dedicated piece.
What This Tells Us About LLM SEO Right Now
The window is still open. For "seo llm agency dubai," the competitive set is small. This is a category where a well-executed, technically rigorous page can reach the top three in a reasonable timeframe. That window will close as more agencies recognise the opportunity. Early movers are establishing citation authority while competitors debate whether it matters.
Google's AI Overview and traditional organic rankings are correlated, not separate. The same page ranking position one to three organically is the page cited in the AI Overview. The underlying signals — topical authority, structured content, clean entity data, relevant inbound links — serve both algorithms. This is one programme, measured in two places.
The llms.txt file is still underused. Of the competitive set we analysed for this keyword, adoption was minimal. It is the fastest single action on this list — low effort, direct signal to AI crawlers, zero cost.
Schema at the page level is where most opportunity sits. Many Dubai agency sites have some schema. Very few have it at the page level, with the right entity types, for the right services. The gap between "has schema" and "has effective schema" is where rankings are won.
Want This Playbook Run on Your Brand?
We audit LLM visibility across ChatGPT and Gemini for UAE and KSA markets, with active querying methodology and GCC-calibrated benchmarks.
See Our LLM SEO Service
.png)
