This is a baseline measurement, not a success story. The brand measured is Vera Bilişim, the company that owns AIShortlist; the measurements were taken in the week of 24 September 2026. There is no before and after comparison yet; it will be added here when there is one.
What did we measure, and how?
Brand: Vera Bilişim (verabilisim.com), a digital marketing and AI consultancy in Istanbul. Engines: ChatGPT, Gemini and Perplexity. Each question was asked to each engine once; a mention counted when the answer named the brand.
We measured the same brand with three different question lists, because which questions are asked is the decision that shapes the result most.
What did the three question lists show?
| Question list | Answers | Mentions | Score |
|---|---|---|---|
| General category questions (8 questions) | 24 | 1 | 4.2% |
| Narrow, service specific questions (10 questions) | 30 | 13 | 43.3% |
| Natural buyer questions for Istanbul (15 questions) | 45 | 3 | 6.7% |
On the narrow questions the gap between engines was large too: Perplexity named the brand on 8 of 10 questions, ChatGPT on 4, Gemini on 1.
On the natural buyer questions the brand was named on three topics: WhatsApp assistants, Instagram message automation and ChatGPT ads. It was never named on ERP, e-invoicing or business intelligence questions; those answers named the large software brands in those fields.
What did we learn from it?
- The score depends on the question list. The same brand can score 4% or 43% in the same week. That is why you approve the questions in the dashboard; nothing is measured without approval.
- Narrow questions that score high feel good but may not be what a buyer really asks. Natural questions that score low show where the work is.
- Engines do not stand in for one another. Being visible on one does not mean you are visible on another; the pages each one cites need a separate look.
- The topics where the brand was named are topics that have a clear, dedicated page on its site. This is an observation, not proof of cause.
What are the limits of this measurement?
- Each question was asked to each engine once. With no repeated runs, the result on any single question is variable.
- The brand owns the product; this is not an independent customer case.
- It is a one-week snapshot. It does not yet show change over time or the measurement after any work.
- The numbers were taken under the conditions described on the methodology page; the answer a user sees in the app can differ.
What is the next step?
The next step is to keep the natural buyer questions fixed and measure them regularly, and to do the jobs the dashboard suggests for the topics where we are not named (missing sub-questions, new pages, list pages), recording the date of each. The comparison will be added to this page when it exists.