RecoSignal vs Profound: Two Different Questions About AI Visibility

The difference in one sentence

Profound measures a brand. RecoSignal measures a business in a place.

That sounds like a slogan, so here is what it means operationally. Profound's unit of analysis is a brand name across a market: how often do assistants mention it, next to which competitors, with what sentiment, citing which sources. That is exactly the right unit if you are a SaaS company, a DTC brand, or a retailer whose customers can be anywhere. RecoSignal's unit of analysis is one business in one city: when a customer inside that city asks an assistant for a recommendation — with the location attached to the question the way a phone attaches it — does the answer contain your name, or the name of a competitor three blocks away.

Neither is a better measurement. They are measurements of different things, and the wrong one produces a number that looks fine and means nothing for your booking calendar.

What Profound does, and does well

It is easy to write a comparison page where the competitor turns out to be worse at everything. That page is not worth reading, and it is not what the evidence supports here. Profound is the most complete enterprise stack in this category, and there are jobs it does that we do not attempt.

• Engine coverage at the top end. Starter covers ChatGPT only; Growth covers three engines; Enterprise reaches up to ten. If your board wants to know what Copilot and Meta AI say about you, that is Profound territory, not ours.
• Daily tracking. Brand mentions are measured continuously, not on the cadence of an audit. For a brand team that has to explain a movement in a weekly meeting, that matters.
• Prompt analytics. Not just whether you were mentioned, but which phrasings of a question produce which answers — the raw material for a content strategy at scale.
• Agent analytics. Visibility into how AI crawlers and agents interact with your site, which is a genuinely different data source from asking the engines questions.
• SOC 2. Unglamorous and decisive: enterprise procurement often cannot sign without it.
• Published pricing: Starter $99/month billed annually (ChatGPT only), Growth $399/month (three engines), Enterprise custom (up to ten engines).

Where Profound is not the right tool

The limitation is not quality — it is fit. Profound was built brand-first, and a single-location dental practice is not a brand in the sense the product means. A practice does not have national share of voice. It has one question, repeated: does the assistant name us when someone nearby asks.

Concretely, the things a local business needs measured are the things a brand tool has no reason to measure. Whether the engines name you when the location is passed as a structured parameter rather than typed into the question. Whether you appear on the urgent question — the chipped tooth tonight — which is a different answer from the researched one. Whether you appear on the insurance question, the device question, the price question. And who, specifically, appears instead of you: not "competing brands in the category," but the four practices your patients are actually choosing between.

That is the honest boundary. If you sell nationally, Profound sees things we do not. If your customers come from a fifteen-minute drive, a national share-of-voice number is not a proxy for anything you can act on.

What RecoSignal measures instead

One RecoSignal audit is 138 live measurements: 30 customer-style prompts on each of ChatGPT, Gemini and Perplexity, with the eight highest-intent prompts asked three times each because these systems are non-deterministic and a single answer is an anecdote. The prompts are written the way customers write, not the way marketers write, and they are spread across five intents. The whole method — including the engine weights, the verdict bands, and what the score cannot prove — is published in full.

The five intents are where the local difference actually lives:

• best_of — "Who is the best implant dentist in [city]?"
• near_me — "Good med spa near me," with the location passed to the engine as a structured parameter where the engine supports one, rather than typed into the question. ChatGPT and Perplexity accept it; Gemini, in our testing, did not act on it, so for Gemini we name the city in the text and say so rather than pretend the three engines were asked identically.
• attribute — the question that filters on a checkable fact: an insurer accepted, a device used, a service offered, a price.
• comparison — "X or Y in [city]: who does both, and what do they charge?"
• urgency — the same-day question: a chipped tooth tonight, a burst pipe, an emergency appointment.
• And the comparison set is your actual neighbours — the practices a patient in your city is choosing between — not a category of national brands.

The data we have that the brand tools do not

The reason we are confident the local question is different is that we measured it, and the answer surprised us. In July 2026 we ran three Dallas markets and counted every source the engines cited behind their recommendations.

In dental, 1,085 of 1,367 citations — 79% — were clinics' own websites. The largest third-party platforms were far behind: Healthgrades 61 citations, Zocdoc 59, Yelp 45 (full benchmark). Medspa came out at 948 of 1,199 own-site citations (benchmark); plastic surgery at 837 of 1,153 (benchmark). Three markets, the same shape: between 73% and 79% of what the engines read about local clinics, they read on the clinic's own site.

That finding has a direct consequence for what you should buy. If the directories were deciding these answers, the right purchase would be listings management. They are not deciding them — your own pages are, by a wide margin. Which means the work that moves a local business is work on its own site: services stated in plain text, prices where a machine can read them, practitioners named with credentials, the insurance question answered on a page instead of on the phone. A brand share-of-voice dashboard will not tell you that, because it is not looking at your street.

It is one metro, three verticals, one measurement window, and we publish the limits alongside the numbers. It is still primary data, and we have not seen the equivalent published by anyone measuring brands.

Price, stated honestly

We are not the cheap option, and the comparison is easy to check.

• Profound: $99/month Starter (billed annually, ChatGPT only) · $399/month Growth (three engines) · Enterprise custom, up to ten engines.
• RecoSignal: free Snapshot · $149 one-time Action Plan · Monitor $79/month · Grow $149/month.
• At the entry level, Profound costs less per month than our one-time plan costs once. If price is the deciding factor, that comparison is not in our favour and we are not going to pretend otherwise.

Which one to choose

The decision is not close once the question is stated properly.

• You are a national or multi-market brand, and the question is share of voice across a category → Profound. Nothing we do substitutes for that.
• You need SOC 2, daily tracking, ten engines, or procurement sign-off → Profound. We do not compete on that axis.
• You are an agency serving enterprise brand teams → Profound, and the agent analytics are a real differentiator.
• You run one location, or a handful, and your customers come from a short drive → RecoSignal. See how an AI visibility audit works, or the vertical pages for dental, medspa and plastic surgery.
• You are an agency whose clients are local businesses → RecoSignal for agencies: white-label reports, and a Snapshot that opens the conversation with a prospect.
• You are both — a national brand with flagship locations that live or die on local demand → the two are complementary, not substitutes. Buy the brand tool for the brand, and measure the locations separately.

What neither tool can promise

No one — not Profound, not us — can guarantee that an assistant will recommend your business. There is no ranking to buy, no #1 position to reach, and the engines change how they answer without notice. Anyone who promises otherwise is either confused or selling.

What measurement honestly does is narrower and more useful: it counts how often you are named across many realistic questions, shows who is named instead, identifies the sources the engines actually cited, and stores the raw answers so any number can be rechecked rather than believed. Closing the specific gaps it finds raises the probability that the next answer includes you. Probability, not certainty, is the whole of the honest claim.

If you want to know where you stand before spending anything, the free AI Visibility Snapshot asks the engines the questions your customers ask and shows the score, the competitors named instead of you, and your single biggest gap.