Profound Prompt Volumes is Profound’s prompt volume data: monthly estimates of how many prompts people ask ChatGPT, Gemini and Perplexity about a topic. Profound Prompt Volumes is genuinely useful when you’re sizing a topic, choosing which prompts to track, or deciding what to write first.
Profound prompt volumes say how often the question gets asked. AI search visibility says what happens when it does.
What are prompt volumes?
Prompt volumes are estimates of how often people ask AI assistants about a topic over a set period, usually a month. They do the job keyword volume used to do: telling you whether demand exists before you spend a quarter chasing it.
Profound counts prompts containing the keywords in the exact order as entered, per its feature page. Evertune clusters prompts into topics instead, because in its analysis over 80% of AI prompts are unique. Its example is the same buyer question asked two ways, “compare Porsche’s electric sedan to Tesla” and “is the Taycan or the Tesla Model S better?”, same intent, completely different wording.
Our read: prompts are conversational, so an exact-phrase count is a floor, and the topic is the honest unit. Either way you’re reading a modeled estimate of demand, not a log of what your buyers typed. Knowing that is what keeps the number useful.
How the prompt volume numbers are made
Prompt volumes are measured and estimated from panels: every number you can buy is a model’s scaling of a real sample, and nobody outside the vendor can see the error bar on it. Hold that in your head and the numbers get a lot easier to use.
Profound’s chain, as its feature page describes it: conversations licensed from multiple double-opt-in consumer panels of real answer engine users, anonymized, aggregated and scrubbed of personal data, then scaled to the full population by statistical modeling that corrects for demographic and geographic biases. The panel is real people asking real questions, currently hundreds of millions of prompts a month from millions of active users.
The data refreshes weekly with less than a week of latency, and the engines covered are currently ChatGPT, Gemini and Perplexity, with DeepSeek and Qwen listed as planned. So treat the number as a planning input, and read demand from your own data first.
Our demand ladder runs: what buyers ask you, on sales calls and in support, plus the queries already reaching you in Search Console. Tool volume comes second, as the wider sample. And the bottom rung is unknown, never zero: a prompt reporting ten or fewer a month is unmeasured, not unimportant.
What prompt volumes are good for
Prompt volumes are used for four planning jobs: sizing a category before you commit budget to it, choosing which buyer prompts to track, spotting a topic that’s growing month over month, and prioritizing what to write so the plan has demand behind it. The first three are read-only, and the number just changes what you believe about the market. The fourth is where teams go wrong, because they rank by raw volume.
But volume alone is a trap. Our position is that you rank topics by estimated traffic value instead, which is click potential, the volume times the click-through you can expect, times the cost per click. The dollar leads, because a dollar number is what buys the meeting with the budget holder.
Let’s say a buyer prompt gets asked 40 times a month at a $13 cost per click, and an informational prompt 400 times at 50 cents. At full click value that’s about $520 a month against about $200. Volume would have ranked them backwards.
Our platform sizes every topic this way before anyone writes anything, with demand, click potential and cost per click side by side. So the priority list is a money list before it’s a content list.
What prompt volumes can’t tell you
Prompt volume data can’t tell you whether the assistant recommends you when the prompt is asked. It doesn’t say which sources the assistant read before answering, and it doesn’t say what to change on your site. Volume tells you how big the room is, and nothing about whether the room picks you.
The number that matters is the share of answers that name you across the prompts you track. That’s what we treat as the visibility number: each prompt asked many times, every answer recorded, your rate next to each competitor’s on the same denominator. It’s a different measurement from a volume estimate, and it’s the one that tells you whether the prompt is worth spending on at all.
Classic demand signals don’t predict AI outcomes either. In our 10x AI Visibility Refresh study, Google rank and AI citation had basically no relationship over 91 days, Spearman rho 0.017 with p at 0.914, and 41 cited pages had zero Google impressions.
And in The Credibility Layer, every recommendation ChatGPT gave came out of a credibility audit of roughly 30 searches, across 200 reasoning sessions, 5,933 logged searches and 1,982 citations. The sources decide the answer, not the prompt’s popularity.
How to put prompt volumes to work
You use prompt volumes to prioritize content in four steps, and each one turns a volume number into work you can defend.
1. Start from prompts your buyers actually type
A prompt belongs on your list when a buyer with budget would type it and it ladders to something you sell: that’s the lane check, and it comes before any volume number does. A prompt that’s mostly job seekers and students can carry a big volume and still be noise. Our platform’s topic intake runs the same lane check first, so a prompt no buyer would type never reaches your list, whatever its volume.
2. Rank the list by value
Rank by estimated traffic value, click potential times cost per click, with your click-through assumption written down where anyone can audit it. Then the ranking is an argument, not an opinion. Our platform’s topic intake runs this ranking before any writing starts, so the loudest voice in the room can’t win on volume alone.
3. Measure the prompts you picked, directly
Run each prompt many times across the assistants and record what comes back, because one probe is basically a coin flip. The rate over many runs is the measurement, and a single answer is an anecdote. Omnipresence asks each tracked prompt many times on ChatGPT, Claude, Gemini and Perplexity, so the share of answers that recommend you is a real rate.
4. Work the gap and re-measure
The cheapest wins are the prompts where you’re named but not recommended. Our method there is to answer the question the heading asks, immediately, in the first paragraph below it, and our agents check each passage’s relevance score against what retrievers measure. Leave a few prompts untouched as controls, then re-run and compare, because the gap between worked and unworked prompts is the proof the lift was yours.
Measuring this in practice
To measure prompts directly in AI search, you ask each buyer question many times and read the rate. That’s the job Omnipresence was built for.
It shows whether ChatGPT, Claude, Gemini and Perplexity recommend you or a competitor on the buyer prompts you pick, the searches behind each answer, and the pages that beat you. A few prompts stay untouched as controls, so the lift you report is provably your own work.
Then it hands the playbook to the agent your team already uses, so the fix gets done where the work already happens, and every change lands on the log next to the number that moved.
Frequently Asked Questions about Profound Prompt Volumes
How much does Profound cost?
Profound doesn’t publish standard pricing. Its pricing page currently shows Enterprise plans at custom pricing, and an FAQ noting that the Starter and Growth plans were retired in September 2026, so the $399 and $499 prices in older reviews are no longer available.
Do prompt volumes exist outside the US?
Yes. Profound currently covers ten countries: the US, Canada, Italy, Brazil, Germany, Australia, Spain, South Korea, France and the UK. Coverage keeps expanding, so check the current list before you size a market.
How many prompts should a team track?
Small and buyer-shaped beats broad. Track enough prompts to cover the buying journey, and few enough that you can work them and re-run them on a schedule. Twenty prompts you re-run weekly tell you more than two hundred you glance at once a month.
Is a high prompt volume good or bad?
Neither, on its own. A high-volume prompt you’re losing is urgent. A high-volume prompt you’re winning is a moat.
And a prompt with volume that your buyers never type is noise, whatever the number says.
Pull the volume numbers when you size a topic, and read them as a range. Then run the prompts you picked and read what comes back. That’s the half that decides anything.