How Perplexity Started Naming Grovia Digital
In September, AI assistants named Grovia Digital in 0 of 525 answers. Four weeks later, Perplexity names us on a real buyer question. What we changed, what it shows, and what it does not prove yet.
Starting point
Between 2 and 9 September 2026 we asked ChatGPT, Perplexity and Gemini 35 buying questions about GEO agencies, five times each: 525 measurements. Grovia Digital was named 0 times and cited as a source once (Gemini, one run). That was our baseline, published in full.
What we changed
- 24 Sep: clearer titles on our service pages and on the agency-selection article. Why: in the baseline, 51.5 % of all cited sources were company websites. Assistants pick the page whose title matches the question. Ours didn't.
- 26 Sep: profile on Clutch. Why: almost a quarter of cited sources (about 23 %) were comparison lists and directories. Grovia Digital appeared in none of them.
- 1 Oct: internal links between our articles and service pages. Why: a technical audit on 24 Sep showed that only 4 of 12 articles linked to our service page, and our Dubai page had a single internal link.
- 3 Oct: one consistent self-description and the same four services across the website, footer, llms.txt and profiles. Why: Search Console showed that about half of the brand impressions for "Grovia" belonged to other companies. Assistants could not tell which Grovia was meant. English pages were aligned to buyers in the GCC, because 11 of 18 Gulf prompt rows had providers that stayed stable across all five runs, against 9 of 45 German rows.
- 6 Oct: listed on DesignRush (Dubai). Why: a third-party source tied to the market where we are based.
The result: Pulse P2, 8 October 2026
90 measurements, three runs per prompt and platform.
- Perplexity names Grovia Digital on the buyer question "Which agency should I hire if my brand doesn't show up in AI assistants?" (asked in German) in 2 of 3 runs, at position 2. It cites our article in all 3 runs.
- On a second question, Perplexity cites the same article in all 3 runs, in the source list.
- ChatGPT and Gemini named us in 0 of 60 measurements.
Why it worked
Perplexity describes us in almost the words of our own article: prompt-based monitoring, work on external sources, a 90-day pilot. That matches the baseline. In the baseline, in about 49 % of provider mentions (in answers with captured sources), the provider's own website was cited in the same answer. The page that answers the question becomes the source. The source becomes the recommendation.
What this does not prove yet
This is one question on one platform. We cannot attribute the effect to a single change. The platforms also move on their own: on control questions we never worked on, new providers appeared for the first time.
What's next
More pages that answer real buyer questions. Next measurements on 22 Oct (P3) and 5 Nov (P4), then a full re-measurement in December.
Raw data
Want the same for your brand?
We start with a 90-day pilot and a documented baseline on a fixed set of buyer questions.