No two AEO tools measure “AI visibility” the same way — Profound counts a weekly mention rate against opt-in consumer-panel conversations, Ahrefs Brand Radar weights impressions by search demand behind 271M+ prompts, and Semrush normalizes everything to a 0-100 index off a 317M-prompt database — so a brand can show “high visibility” in one dashboard and “low” in another without either tool being wrong.
That’s the first thing worth knowing before buying an AEO (Answer Engine Optimization) or “AI visibility” platform: these products don’t share a metric standard the way Google Analytics and Adobe Analytics roughly agree on what a “session” is. Each vendor built its own prompt corpus, its own brand-extraction logic, and its own definition of “share of voice.” Below is what each platform’s own documentation says it measures, not what its marketing page implies.
The five platforms and what they actually track
Profound
Profound calculates a Visibility Score as the percentage of tracked responses that mention a brand, recomputed on a rolling weekly basis (Profound, 2026). Its prompt set — what it calls “Prompt Volumes” — comes from licensed, double opt-in consumer panels of real AI users rather than search-volume data or synthetic queries (Profound, “Introducing the Profound Index,” 2026). Pulling model coverage directly from Profound’s own API on September 4, 2026 shows eight tracked surfaces for a standard account: ChatGPT, Google AI Mode, Google AI Overviews, Google Gemini, Grok, Meta AI, Microsoft Copilot, and Perplexity — Claude is notably absent from this live pull, though third-party comparison posts describe Profound covering up to 10-11 surfaces on higher tiers. That gap between marketing copy and what an API call actually returns is exactly the kind of discrepancy this site tracks — see how to evaluate LLMs for the same pattern in benchmark leaderboards.
Ahrefs Brand Radar
Brand Radar builds its prompt library from Google’s “People Also Ask” corpus plus Ahrefs’ keyword database, expanded with a “Fanout” system into related sub-questions — the vendor calls these “search-backed prompts, not synthetic ones” (Ahrefs, 2026). Its Share of Voice metric is the percentage of brand impressions out of total impressions among responses that mention any tracked brand, and impressions are weighted by the search demand behind the prompt that surfaced them — so ranking in ten low-volume questions doesn’t count the same as ranking in one high-demand one. A live schema pull from Ahrefs’ own API on September 4, 2026 lists ten selectable data sources for Brand Radar: chatgpt, google_ai_overviews, google_ai_mode, gemini, perplexity, copilot, claude, grok, and two keyword-derived variants for AI Overviews and AI Mode — more than the “six AI surfaces” figure that circulates in third-party reviews of the product.
Semrush AI Visibility Toolkit
Semrush normalizes its output to a 0-100 AI Visibility Score built from a prompt database it describes as 317M+ AI queries, drawing on Google AI Overviews, AI Mode, Gemini, and ChatGPT search mode (Semrush Knowledge Base, “Where does the data in Semrush’s AI Visibility Toolkit come from?,” 2026). Brand identification runs through a proprietary extraction system Semrush says can distinguish, for example, Tesla the company from Nikola Tesla the person — a real disambiguation problem for any brand with a common name. Data refreshes daily on a rolling basis rather than the weekly cadence Profound uses.
Peec AI
Peec’s documentation defines its atomic unit of measurement as a “chat”: one prompt, run against one model, in one location, on one day, producing one response (Peec AI Docs, “Visibility,” 2026). Visibility is the percentage of chats mentioning a brand at all; Share of Voice is that brand’s mentions as a percentage of all tracked-brand mentions within those chats — a distinct calculation from Ahrefs’ impression-weighted version. Peec also publishes a 0-100 sentiment score per brand per engine, which none of the larger suites expose as a standalone number in the same way.
Otterly.AI
Otterly defines Share of AI Voice as the percentage of citations a brand owns versus competitors across a defined prompt set, tracked daily across ChatGPT, Perplexity, Google AI Overviews, AI Mode, Gemini, and Copilot (Otterly.AI, 2026). Its KPI set is the widest of the group at a glance — brand mentions, Share of Voice, average brand position, brand coverage, domain citation, domain coverage, and a composite Brand Visibility Index — but it is also the only tool here where citation-based metrics, not raw mention rate, are the primary lens.
Where the metrics genuinely disagree
Two disagreements matter more than the rest. First, “share of voice” means different things depending on the vendor: Ahrefs weights it by search demand, Peec weights it by raw mention count within tracked chats, and Otterly frames it around citations rather than mentions — a brand cited once in a low-traffic prompt and mentioned five times in high-traffic ones will not rank the same way across all three. Second, prompt sourcing splits into two camps: Ahrefs and Semrush derive prompts from real search-query and keyword data, while Profound licenses actual consumer-panel conversations — meaning Profound’s numbers reflect what people actually typed into a chatbot, while the SEO-platform tools reflect what people search for, extrapolated into AI answers. Neither approach is strictly more accurate; they’re measuring adjacent but different behaviors, which is why cross-checking a single brand across two of these tools before reporting a number externally is the safer default — the same cross-check discipline this site applies to benchmark scores in How to Evaluate LLMs.
| Platform | Core metric | Prompt source | Surfaces tracked (per vendor/API) | Refresh |
|---|---|---|---|---|
| Profound | Visibility Score (% of responses mentioning brand) | Licensed consumer-panel conversations | 8 confirmed live via API (Sept 2026); vendor claims up to 10-11 | Weekly |
| Ahrefs Brand Radar | AI Share of Voice (demand-weighted impressions) | Google PAA + keyword DB + Fanout expansion | 10 data-source options via API (incl. 2 keyword-derived) | Scheduled snapshots |
| Semrush AI Toolkit | AI Visibility Score (0-100, mention-based) | 317M+ query database | 4 named (AIO, AI Mode, Gemini, ChatGPT), “more coming” | Daily, rolling |
| Peec AI | Visibility % / Share of Voice per “chat” | User-defined prompt sets | Multiple, incl. 8-engine studies cited by vendor | Daily per engine |
| Otterly.AI | Share of AI Voice (citation-based) | User-defined prompt sets | 6 named (ChatGPT, Perplexity, AIO, AI Mode, Gemini, Copilot) | Daily |
Pricing tiers roughly map to buyer type
Self-serve entry points cluster around $29-99/month — Otterly’s Lite plan starts at $29/month, and Peec AI’s self-serve tier sits near $89-99/month, according to comparison research aggregating current list prices (Surmado, 2026). Profound and enterprise-tier competitors are typically sold on annual contracts without public list pricing. Ahrefs and Semrush bundle AI visibility into existing SEO suite subscriptions rather than pricing it standalone, which is the practical reason teams already paying for one of those platforms tend to start there before evaluating a dedicated AEO tool.
What this means for measurement discipline
None of these five tools should be treated as a single source of truth for “AI visibility,” for the same reason no single benchmark should be treated as a single source of truth for model capability. If a number is going into a board deck or a press claim, pull it from at least two platforms with different prompt-sourcing methods — a consumer-panel tool like Profound and a search-derived tool like Ahrefs or Semrush — and report the range rather than a single figure. Where the tools’ numbers should feed directly into action is technical: if a brand’s visibility score is near zero everywhere, the more useful next step is checking whether AI crawlers can even reach the site, covered in AI Crawler Permissions, before assuming the content itself is the problem. For a structured, tool-agnostic way to audit visibility without a subscription, see the DIY AI Visibility Audit. And because these platforms measure citation and mention behavior differently by design, understanding how ChatGPT, Perplexity, and Gemini choose citations differently explains a good chunk of why the same brand scores unevenly across dashboards. All of this sits under the broader question of how AI assistants decide which brands to recommend in the first place — the tools in this piece are instruments for observing that process, not the process itself.
FAQ
Do any two AEO tools produce the same visibility score for the same brand?
No. Because each platform uses a different prompt corpus (consumer-panel conversations, search-derived keywords, or user-defined prompt sets) and a different share-of-voice formula (demand-weighted, raw-mention, or citation-based), the same brand can score differently across tools even when tracking the same AI surfaces on the same day. Treat any single tool’s number as directional, not absolute.
Which AEO tool tracks the most AI platforms?
Based on live API/data-source options pulled September 4, 2026, Ahrefs Brand Radar exposes the most data-source options (ten, including ChatGPT, Gemini, Perplexity, Copilot, Claude, Grok, and two Google AI variants), followed by Profound’s eight confirmed surfaces on a standard account. Vendor marketing pages sometimes cite higher surface counts than what a given account’s API actually returns, so it’s worth checking your own plan’s data sources directly rather than relying on the comparison page.
Should a brand buy a dedicated AEO tool or use the AI visibility features already in Ahrefs or Semrush?
Teams already paying for Ahrefs or Semrush get AI visibility bundled into their existing subscription, which is a reasonable starting point for a first read. Dedicated platforms like Profound, Peec AI, or Otterly.AI generally offer deeper prompt customization, per-engine sentiment scoring, or panel-based (rather than search-derived) prompt data — worth the added cost mainly once a team needs to act on AEO numbers rather than just monitor them.
Last updated September 4, 2026. This page is refreshed as benchmarks and scores move.