FLPL
Dibyendu MondalShort answer
Write 20 questions your buyers ask, run each one in ChatGPT, Perplexity and Google's AI from a clean session, and record who gets named. Count your mentions against your top competitor. Repeat monthly with the same questions so you're measuring change, not noise.
Step 1: write the questions your buyers ask
Not keywords, questions. Think about what a buyer types when they don't know your name yet:
- "best [category] for [type of company]"
- "[competitor] alternatives"
- "[category] that integrates with [tool they use]"
- "cheapest [category] for [size]"
- "[job to be done] software"
Aim for 20. Use your sales calls, support tickets and the questions prospects actually asked you. Don't include your brand name: you want to know if you show up when nobody asked for you.
Step 2: run them cleanly
- Use a logged-out or private browser session where you can, so your own history doesn't skew the answers.
- Ask each question exactly as written, once per tool: ChatGPT, Perplexity and Google (AI Overview or AI Mode).
- Note the date. Answers change, so a dated record is what makes month-to-month comparison possible.
Step 3: record three things per answer
- Were you named? Yes or no.
- Who was named? Every company, in order.
- Which sources were cited? Perplexity and Google show them, and ChatGPT often does when it searches. These are the pages you need to be on.
A simple spreadsheet works: one row per question, one column per tool.
Step 4: read the results
With 20 questions across 3 tools you have 60 answers. Count how many name you and how many name your strongest competitor. That gap is your starting point.
Then look at the cited sources. If the same three comparison sites keep appearing and you're not on them, you've found your first job.
Step 5: repeat monthly
Use the same questions, in the same way, around the same date each month. One run tells you where you stand. The trend tells you whether your work is moving anything.
Don't want to run it yourself? The free FLPL check does steps 1 to 4 for you and ends with your top five fixes. See a sample report.
Questions people also ask
Why do I get different answers each time?
AI answers vary between runs, users and days. That's why you track a fixed set of questions over time and look at the trend instead of any single answer.
Should I include my brand name in the questions?
Not for this check. Branded questions tell you whether the AI describes you correctly, which is useful too, but they don't tell you whether you're recommended to buyers who haven't heard of you.
How many questions are enough?
Twenty is a practical start. Fewer than ten and one odd answer swings the result; more than fifty takes long to run by hand every month.