Free tool
Buyer question generator
Enter your domain or describe your business and get ten unbranded questions real buyers would ask an AI assistant in your category. It runs a small language model on our server, so runs are limited per day. The questions matter because they are the ones worth tracking: ask an assistant about your brand by name and you will always look visible.
Branded questions fake your visibility score
Ask ChatGPT 'is Acme any good?' and Acme will dominate the answer, because the model is completing a prompt that already contains the brand. Every vendor demo that shows a brand shining on its own name is showing you this artifact. It measures nothing about discovery.
Discovery happens on unbranded questions: 'best appointment software for small clinics', 'how do I get more reviews for a local business'. Those are the moments an assistant hands out recommendations to buyers who have never heard of you, and where a rival being named in your place is invisible revenue loss. The ten questions this tool generates are of that kind, phrased the way people actually talk to assistants: conversational, specific and tied to a situation.
How to use the ten questions
First, audit. Paste each question into ChatGPT, Perplexity and Gemini and note which brands come back. That half hour usually settles whether you have an AI visibility problem more convincingly than any report could.
Second, treat lost questions as content briefs. Each question where rivals get named and you do not is a page you have not written, or a page that fails to answer directly. Third, track over time, because answers shift as models and their sources update. A one-off check is a snapshot; the trend is the signal.
Honest limits
A small model generates plausible category questions, and plausible is the ceiling. It has never heard your sales calls, read your search query data or seen your support tickets, which is where the truest buyer questions live. Cross-check its output against what prospects ask on calls, what appears in Google's People Also Ask boxes, and what your category's community threads keep debating.
Runs are limited per day because each one costs real inference money on our side. If you want a maintained question set with tracked answers across engines, that is the paid job: Reachroller tracks whether ChatGPT and other assistants name your brand on your buyer questions, free for 3 days and from $29 a month.
Frequently asked questions
Why unbranded questions only?+
Because branded prompts contaminate the measurement. A question containing your brand name pulls your brand into the answer almost every time, which makes any visibility score built on it meaningless. Unbranded category questions are where assistants choose freely among competitors, so they are the only honest basis for tracking whether you get recommended.
How many of the ten should I track?+
Five to ten is plenty to start. Pick the ones closest to a purchase decision in your category and check them on a fixed cadence, weekly works well, across the same engines each time. Consistency beats volume here, since the point is spotting movement after you publish or a model updates.
Why is this tool limited to a few runs per day?+
Each run calls a language model on our server, and inference costs actual money per request. Rather than gate the tool behind an email form, we cap daily runs and keep it free. If you burn through the cap, come back tomorrow, or generate questions for one business and spend the day auditing them.
How accurate are the generated questions?+
Treat them as strong drafts. The model knows how buyers phrase questions in most categories, and it misses the specifics only you can know: your segment's jargon, objections from sales calls, regional quirks. Verify against real conversations and community threads, swap in the questions your prospects genuinely ask, and keep the rest.
Tools find the gaps. Tracking closes them.
Reachroller asks the questions your buyers ask and shows which brands ChatGPT names. Three days free, no card.
Check my brand free