Prompt tracking

AI prompt tracking for the questions buyers actually type

Every number on every other page depends on this one choosing well. We suggest a prompt set from your category and competitors, put real search volume beside each suggestion so you can sort by demand, then ask the ones you keep across all six engines on a fixed schedule.

Start 7-day free trialAll seven features

Card required, nothing charged for 7 days

Sample · “which CRM should a small law firm use”

One check

1 / 1

Shown as a fraction. A single answer is a draw, not a rate.

14 recent runs

71%

10 of 14 answers named you. Now it is a rate.

Same question, same engine, one answer per day. Failures on our side are excluded from the denominator.

Illustrative figures from a demo workspace, not a customer's data.

From a blank list to a tracked set

Four steps, and the second one is the part most tools leave you to guess.

  1. 1

    Suggest

    A candidate set generated from your category, your brand and the competitors you name.

  2. 2

    Price

    Real monthly search volume beside each candidate, so you can sort by demand rather than instinct.

  3. 3

    Choose

    Accept, edit or write your own. Between 2 and 15 depending on plan.

  4. 4

    Track

    Every chosen prompt, on all six engines, on a fixed schedule, from the day you add it.

Demand, not instinct

Search volume beside every suggestion

The usual way to pick prompts is for somebody in marketing to write twenty questions they think buyers ask. Some are right, some are what the team wishes buyers asked, and there is no way to tell which is which. We price each candidate against real Google search volume first, so the list you approve is sorted by how many people ask it. Prompts you write yourself are tracked exactly the same way, and everything is measured in the market you choose.

Priced before you choose.
Volume is looked up live and cached for a week, so the list is current without being billed twice.
Your market, your demand.
Volume is fetched for the market the workspace measures, so demand and measurement describe the same place.
Degrades honestly.
If the volume vendor is down you get unlabelled suggestions, not invented numbers.
Overview · last 30 days · 1,240 answers
EngineNamedCitedn
ChatGPT47%63%30
Claude30%37%30
Gemini57%70%30
Perplexity40%80%30
Google AI Overviews20%50%30
Google AI Mode3/44/44

Google AI Mode shows a fraction: below five answers a percentage would be a guess.

Illustrative figures from a demo workspace, not a customer's data.

Nobody can tell you real AI prompt volume

Including us. What we show is search volume, and the distinction is worth being pedantic about.

Several tools in this category advertise prompt volumes as though they were measured. They are not. The assistants do not publish how many times a question was asked, and there is no panel watching enough conversations to infer it. Every figure on the market is a model built on search data, and the honest versions say so.

Ours is search volume for the phrasing, presented as search volume. The assumption is that someone who types a question into Google is much the same person who types it into ChatGPT, which is reasonable and is still an assumption. It is a good way to rank candidates against each other and a bad way to forecast traffic. The same refusal to dress up an estimate runs through how we report visibility and what we will not claim in our methodology.

  • Labelled as search volume, never as prompt volume, because that is what the number is.
  • Missing rather than guessed. A candidate with no volume data shows no volume, not a zero and not an estimate.
  • History starts when you add it. A new prompt has no past, here or anywhere else.

Common questions about AI prompt tracking

What people ask before they commit to a prompt set.

What is prompt tracking?
Deciding which questions your buyers ask an assistant, then asking those same questions on a schedule and recording what comes back. The prompt is the unit of measurement here, the way a keyword is the unit in search, and choosing the wrong set of prompts makes every number downstream describe a conversation nobody is having.
How do I know which prompts to track?
We suggest a set from your category, brand and competitors, and put a real monthly search volume beside each suggestion so you can sort by demand instead of guessing. You accept the ones that look right and edit or add your own. The suggestions are a starting point, not a decision.
Where does the volume figure come from?
Google search volume for that phrasing, looked up live and cached for a week. It is a proxy: it tells you roughly how many people ask that question, on the assumption that someone typing it into Google is much the same person typing it into ChatGPT. It is not a count of how many times the prompt was sent to an assistant.
Can anyone measure real AI prompt volume?
No, and treat any tool claiming otherwise with suspicion. The assistants do not publish prompt counts, and there is no panel that observes them at scale. Estimates in this category are models built on search data, which is what ours is too — we would rather call it search volume, which is what it is.
How many prompts can I track?
Between 2 and 15 depending on plan, each one checked on all six engines on a fixed schedule. If you need more than 15, email us and we will talk about it rather than invent a number on the pricing page.
Can I change my prompts later?
Yes, at any time. Measurement for a new prompt starts the day you add it, because nobody can retrieve what an assistant said last month. That is worth knowing before you swap out a prompt you have six months of history on.

The rest of what we measure

Six more, all on every plan.

Pick your prompts and start tracking

Tell us your category and we will suggest a set with volumes attached. Card required, nothing charged for 7 days.

Start 7-day free trialSee pricing