DAFZ (Dubai Airport Free Zone) came into this program with strong classic SEO but zero visibility inside the AI answer engines that were starting to sit above search results for exactly the queries that matter, "airport free zone," "companies in DAFZ," "DAFZ industrial park," and dozens of adjacent setup and compliance questions. This is the methodology run over 8 weeks, and the numbers behind it.
Starting point
DAFZ's content was written for humans skimming a results page, long paragraphs, context spread across several sentences, no explicit schema. Good enough for classic SEO, invisible to an AI model deciding what to cite in a synthesized answer.
What changed, week by week
Weeks 1–2: Audited every high-intent page against a fixed set of real questions ("what companies operate in DAFZ," "how do I set up a company in DAFZ," "DAFZ vs other UAE free zones"), run manually across Google SGE, ChatGPT, and Perplexity to see who was currently getting cited, and why.
Weeks 3–5: Rewrote priority pages so every section opened with a direct,
self-contained answer sentence, added FAQPage and Organization schema, and
built out the topical depth needed to fully answer each question rather than partially cover it.
Weeks 6–8: Extended the same structure to DIEZ in parallel, and began tracking citation appearances and unique AI-driven search volume as a standing KPI, not a one-time audit.
Why this matters more than the headline number
The keyword footprint also grew in the process, 36 keywords in the Top 3, 66 in the Top 10, 83 in the Top 100, which confirms GEO and SEO aren't competing investments. The same structural discipline that makes a page citable also makes it rank better. The visibility jump (31.61% to 37.67%) happened alongside the GEO gains, not instead of them.
The pages that got cited most were never the longest, they were the ones that answered the actual question in the first two sentences and then kept going.
What I'd tell another marketer running this for the first time
Don't wait for a platform to give you a "GEO traffic" report, there isn't one yet. Sample manually, on a fixed query set, on a fixed cadence, across all three major engines. Track it like you'd track rank position: same queries, same interval, logged consistently, so the trend is real and not noise from query variance.