Search has expanded to AI. The brands it recommends win the customer.

Prompt Research
for AI Search Visibility

Prompt research identifies the exact questions your buyers ask ChatGPT, Claude, Gemini, and Perplexity before they buy. We analyze your sales calls, support tickets, Reddit threads, and search data, then rank every prompt by purchase intent and whether your brand shows up in the answer today.

Book a Call

The Prompt Set

Every question your buyers ask AI, ranked by purchase intent

Prompt set

Top prompt

best corporate card

Intent

High

Your Visibility Baseline

Where you show up today, and who shows up instead

Named in answers

142 prompts tested

Competitors 58%

You 26%

Intent stage

Every prompt tagged problem-aware, solution-aware, or ready to buy, so you know which answers to win.

Competitor presence

Which brands the model recommends instead of yours, and how often, prompt by prompt.

Citation sources

The exact URLs each model pulls from to build its answer, ranked by citation frequency.

Volume and trend

Estimated monthly prompt volume and direction, so you prioritize by where demand is heading.

We start with your actual conversations

Sales call transcripts, support tickets, onboarding notes, and the subreddits and forums where your category gets discussed. These are the sources where buyers phrase things the way they'd phrase them to an LLM, which is nothing like how they type into Google.

Then we cluster prompts, not keywords

The same buying question shows up in a dozen phrasings. We group them into clusters and score each one on purchase intent and estimated volume, so you end up with topics worth owning instead of a keyword list.

Finally, we test every cluster against the models

We run each prompt through ChatGPT, Claude, Gemini, and Perplexity, log whether you appear, which competitors appear instead, and which URLs the model cited to build its answer. That's your baseline, and it's what every later optimization gets measured against.

Every model your buyers use

What is prompt research?

Prompt research identifies the exact questions buyers ask ChatGPT, Claude, Gemini, and Perplexity when they're evaluating a purchase in your category. It replaces keyword research for AI search: instead of a list of search terms, you get the conversational questions people actually type, ranked by purchase intent and mapped against whether your brand appears in the answer.

How is prompt research different from keyword research?

Keywords are fragments people type into Google. Prompts are full questions people ask a model, often several sentences long and layered with context about their situation. A buyer who searches "pilates studio LA" will ask an LLM something closer to "I'm new to pilates and have a bad back, what studio in Los Angeles should I try first." Keyword tools don't capture that phrasing, and optimizing for the fragment doesn't win the answer.

Where do the prompts come from?

Your sales call transcripts, support tickets, onboarding notes, and the forums and subreddits where your category gets discussed. Those sources capture how buyers phrase questions in conversation, which is much closer to how they phrase them to a model. We supplement with search data and query expansion, but customer language is the foundation.

How many prompts do you find?

It depends on category breadth. A narrow B2B product might yield 40 to 80 meaningful prompts. A consumer brand with a wide use case can run into several hundred. What matters more than the count is how many are ranked high-intent, since those are the ones worth building for.

How do you measure purchase intent?

Each prompt is scored on where the buyer is in their decision. Problem-aware prompts describe a symptom without naming a solution. Solution-aware prompts compare approaches or categories. Ready-to-buy prompts name products, ask about pricing, or request a recommendation. Ready-to-buy prompts convert at the highest rate, so they get prioritized even when volume is lower.

Can't I just ask ChatGPT what people ask about my category?

You'll get plausible-sounding prompts, but they're generated, not observed. A model asked to guess buyer questions returns the generic version of your category rather than the specific language your buyers use, and it can't tell you which prompts you're currently losing or who's winning them instead. The value is in the observed data and the baseline test, not the list.

What do I do with the prompt set once I have it?

It becomes the target list for everything downstream: which pages to build, which questions to answer on your site, which sources to get cited by, and what to measure against. The baseline test tells you which prompts you're already losing, so you know where the gap is before spending on content.

Prompt Research
for AI Search Visibility

The Prompt Set