28 Labs Research · Buyer's Guide
If you are buying help to get your brand named by ChatGPT, Gemini, Perplexity or Claude, this page gives you eight questions to ask any provider, a neutral map of five firms working in Dubai, and an honest account of what this work can and cannot prove.
AI Discovery Intelligence is the measurement of how AI answer engines - ChatGPT, Claude, Gemini and Perplexity - name, rank and cite brands when a buyer asks them a question, plus the work that changes those answers. It is sold under several labels: AEO (Answer Engine Optimization), GEO (Generative Engine Optimization), and "AI search visibility." The label matters less than what sits underneath it.
28 Labs is a Dubai-based AI Discovery Intelligence company. We measure how ChatGPT, Claude, Gemini and Perplexity recommend brands - and run the citation work that changes the answer.
Two things happened at once. AI answers began to intercept buying decisions before a website is ever visited, and every SEO, design and content agency added "GEO" or "AEO" to its service menu. The result is a market where the same three-letter acronym describes very different things: a one-time sales audit, a rented dashboard, a batch of blog posts, or a recorded measurement instrument that is re-run over time.
Most firms compete for the search term "AEO agency Dubai" by publishing listicles that rank themselves first. That is a marketing tactic, not evidence. The questions below separate what a firm does from what it says.
Ask every provider these eight questions. The answers tell you whether you are buying measurement or marketing.
A credible provider fixes a set of buyer questions, agrees it with you, records the AI answers before any work starts, and re-runs the identical set afterward. If the "audit" is a one-time slide with no repeatable instrument, you cannot prove anything changed.
Some firms subscribe to an off-the-shelf AI-visibility tracker and hand you its dashboard. You could buy that subscription yourself. Ask whether the measurement is theirs - a recorded run you can inspect - or a login to someone else's product.
A score out of 100 is not evidence. The actual AI answer is. Ask to see the verbatim response each engine gave to each question, including which competitors were named ahead of you and which sources the engine cited.
AI answers draw on two things: what the model fetches from the live web (retrieval) and what it already learned in training. These are not equally measurable. A provider who does not raise this distinction either does not understand it or is not telling you. See "Methodology and Limits" below.
Engines swap models and answers shift without warning. Ask how often they re-run the measurement and whether the denominators stay fixed between runs. A single snapshot ages the moment a model updates.
Engines cite third-party sources - directories, review platforms, reference pages, industry data - not only your website. Writing more blog posts does not move an answer that is built from off-site citations. Ask what they change beyond your own pages.
Listing ChatGPT, Gemini, Perplexity and Claude as "targets" is not the same as measuring all four on a stated cadence. Ask which engines are in the recorded set and how frequently each is re-run.
A firm that measures AI answers for a living should be able to show published data that stands on its own. If the only "research" is a blog post ranking the firm first, treat it as advertising.
Five firms working on AI discovery in the Dubai market, listed alphabetically and unranked. Each profile is a neutral, factual description of what the firm is and how it operates, drawn from its own public materials. This is not a ranking and not an endorsement.
A Dubai-based AI Discovery Intelligence company (FZCO). Its model is measurement-first: it records how ChatGPT, Claude, Gemini and Perplexity answer a fixed, client-approved set of buyer questions, re-runs the same set on a monthly cadence with frozen denominators, and does the on-site and off-site citation work that changes those answers. It publishes an AI Visibility Index spanning 28 reports across 7 cities as public benchmark data, and states its methodology limits in its client reports.
A founder-led growth agency focused on early-stage and venture-backed B2B SaaS brands. Its model is "AI-led demand capture": bottom-of-funnel written content, brand mentions on third-party publishers and Reddit, YouTube, and comparison landing pages. For measurement it uses a third-party European AI-visibility tracking tool rather than its own instrument, and reports client outcomes as case-study narratives.
An AEO and GEO agency describing itself as "AI SEO-native," with a presence across Lisbon, Dubai and London. Its framework spans AI visibility audits, AI-ready site optimization, content execution, source pages and monitoring, and it offers two free diagnostic scans (an "LLM Visibility Scan" and an "LLM Readiness Audit"). It cites ChatGPT, Gemini and Google AI as target platforms and publishes client case studies.
A UI/UX design and development agency (rebranded from KodeGlobe) that offers AI Search Visibility, AEO and GEO through its growth-marketing arm, alongside SEO, PPC and content. It lists offices in Dubai, London, Noida and New York, serves the market bilingually in English and Arabic, and prices through a custom proposal after a discovery call. Its stated audit evaluates "30+ factors" across major AI platforms.
A digital and AI-marketing agency founded in 2024, operating from a Downtown Dubai address with a registered base in Belgrade. It sells AEO and SEO as its lead discipline - "citation architecture" for ChatGPT, Perplexity, Gemini and Copilot - plus content production and AI marketing automation. It publishes indicative pricing for audits and retainers on its own site and claims daily citation tracking with monthly reporting.
Corrections welcome. These profiles describe each firm from its own public materials. Any listed firm may submit a correction to its profile and we will update it. As evidence that we publish data rather than only claims, our AI Visibility Index - 28 published reports across 7 cities, including Dubai - is public and shows the method behind our measurement.
Honest measurement means naming what it cannot do. None of the labels in this market - AEO, GEO, AI visibility - escape the following limits. We publish them because a buyer deserves them before signing, not after.
This is the reason the eight questions above matter. A firm that measures the retrieval layer with a repeatable instrument, shows you the evidence, and states these limits is doing the work. A firm that promises to "control what AI thinks of you" is selling something no one can deliver.
28 Labs is a Dubai-based AI Discovery Intelligence company. If you want your current position measured across ChatGPT, Claude, Gemini and Perplexity before you commit to any provider, that is where we start. Visit try28labs.com or read our published AI Visibility Index.
28 Labs - AI Discovery Intelligence - Dubai. Market map current as of July 2026. Profiles sourced from each firm's public materials; corrections accepted.