More and more buyers skip Google and ask an AI assistant: "What is the best CRM for a small business?". If the answer doesn't name you, you are invisible to that buyer. This is what people now call GEO (generative engine optimization) or AEO.
Here is a simple, repeatable way to measure it, with real numbers from a test I ran in October 2026.
1. Write the questions your buyers really ask
Not "What is Pipedrive?" (you'll always be mentioned), but category questions:
- What is the best CRM for a small business?
- Which CRM is best for a sales team of 10 people?
- Pipedrive vs HubSpot for startups?
10–30 questions per market is a good start. Do it per country and language: answers differ.
2. Ask every engine, and record four things
For each answer:
- Is your brand mentioned?
- At which position? Being 4th in a list of 5 is not the same as being 1st.
- Which competitors are named, and in what order.
- Which websites are cited as sources. ChatGPT, Gemini and Perplexity link the pages they used. Those review sites and comparison articles are where you need to appear.
3. Turn it into three numbers
- Visibility %: share of answers that mention you.
- Average position when you are mentioned.
- Share of voice: your mentions ÷ all brand mentions (you + competitors).
A real test
Brand: Pipedrive. Competitors: HubSpot, Salesforce, Zoho, monday.com, Freshsales.
2 questions × 4 engines (ChatGPT with web search, Gemini, Perplexity, Claude) = 8 answers:
| Metric | Result |
|---|---|
| Visibility | 100% (8/8 answers) |
| Average position | 2.1 |
| First place | 3 of 8 |
| Pipedrive's own site cited as a source | 3 of 8 |
| Share of voice | HubSpot 20.5%, Pipedrive 20.5%, Zoho 20.5%, Salesforce 15.4% |
| Most cited domains | salesforce.com, techradar.com, pipedrive.com, zite.com… |
Takeaways: Pipedrive is always there but comes first in only 3 answers out of 8, HubSpot and Zoho are named just as often, and third-party review sites like techradar.com shape the answers. That's the to-do list.
One answer proves little: AI answers change between runs. Track the averages weekly.
Automating it
By hand, this is fine for 5–10 questions. I got tired of copy-pasting, so I built a tool on Apify (disclosure: I'm the developer): AI Brand Visibility Tracker. You give it your brand, your competitors and your questions. It asks ChatGPT, Gemini, Perplexity and Claude and returns one row per answer, plus the summary above. It costs about $0.03 per answer, with no subscription, and you can schedule it weekly.
Example input:
{
"brand": "Pipedrive",
"brandDomain": "pipedrive.com",
"competitors": ["HubSpot", "Salesforce", "Zoho"],
"prompts": ["What is the best CRM for a small business?"],
"engines": ["chatgpt", "gemini", "perplexity"]
}
Whatever tool you use, the method is what matters: fixed questions, several engines, averages over time, and work on the sources the AIs cite.
Top comments (0)