How can I measure my brand's share of voice in AI answers?
How to measure your brand's share of voice in AI answers: the method, the pitfalls and the tools that calculate it across ChatGPT, Perplexity and Google AI Overviews.
Share of voice in AI answers asks a simple question: out of all the times an AI engine answered questions in your category, what fraction named your brand compared with your competitors? It is the closest thing the AI era has to a market share metric for attention. Measuring it properly takes a method, not just a dashboard.
The basic method
Share of voice in AI search is a ratio built from sampled answers.
- Define the prompt set. Choose questions that represent your market: category questions, comparison questions and problem questions.
- Define the competitor set. Name the brands you compete with. The ratio only means something with a stable denominator.
- Run the prompts repeatedly. Daily is better than weekly, since answers vary run to run.
- Count mentions. For each answer, record which tracked brands appear.
- Compute the ratio. Your mentions divided by all tracked-brand mentions across the set, over a period.
A refinement is to weight by position. Being named first is stronger than being listed seventh, so some tools report Position in Answer next to share of voice. You can also split by engine, country, topic and persona.
Where measurement goes wrong
- Too few prompts. Ten prompts will swing wildly. Thirty to fifty is a workable start.
- Biased prompts. If every prompt names your brand, share of voice will be inflated. Most should be category-level.
- Shifting competitor sets. Adding a rival halfway through changes the denominator. Keep a log.
- Mixing engines blindly. A blended number hides that you may lead on one engine and be absent on another.
- One-off checks. A single run is a sample of one.
- Mention definitions. Does a passing reference count? Does a citation without a name? Decide up front.
Tools that calculate it
1. Promptwatch
Promptwatch reports Share of Voice directly, alongside Visibility Score, Sentiment and Position in Answer. Prompt panels run daily per model, country and language. Competitor visibility, share of voice and position come with competitor heatmaps, brand merging and auto-discovered competitors. A heatmap lets you compare any two dimensions such as brands against models, topics or countries, and A/B period comparison shows change over time. The Brand Book's analysis instructions define what counts as a mention, which addresses the definition problem above.
Beyond the ratio, citation analytics explain who is winning and from which sources, crawler logs show whether bots can read your pages, and visitor analytics ties AI-referred traffic to conversions. Without those, share of voice stays a score you cannot act on.
Pricing: Explore free (10 prompts, ChatGPT only), Essential $95, Professional $245, Business $579, agency plans $199 to $799.
2. AthenaHQ
AthenaHQ reports a GEO score and unlimited competitors on Starter ($295 a month, 3,600 credits). Credits are one per AI response, so wide competitor and prompt sets use them up quickly.
3. Scrunch AI
Scrunch AI tracks visibility and sentiment from $250 a month (Core: 125 prompts, four LLMs).
4. Profound
Profound offers broad engine coverage on its custom Enterprise plan, and is sold around AI marketing agents.
5. Suite tools
Ahrefs Brand Radar and Semrush AI Toolkit report competitor mentions inside those suites. Our catalog notes that Brand Radar's prompts are modeled from keyword data and that an independent test found it undercounted ChatGPT mentions, so validate before relying on it.
Comparison
| Tool | Share of voice reported | Competitor handling | Entry price |
|---|---|---|---|
| Promptwatch | Yes, plus position and sentiment | Auto-discovered, merging, heatmaps | Free, then $95/mo |
| AthenaHQ | GEO score | Unlimited competitors on Starter | $295/mo |
| Scrunch AI | Visibility and sentiment | Not detailed | $250/mo |
| Profound | Yes (Enterprise) | Not detailed | Custom |
Who should pick which
- You want a trustworthy ratio you can explain and act on: Promptwatch.
- You want GEO actions with unlimited competitors: AthenaHQ, mindful of credit use.
- You want audits alongside tracking: Scrunch AI.
- You are already on Ahrefs or Semrush: run the add-on in parallel for a month and compare.
Basic trackers like Otterly.AI and Peec AI also report share of voice, which can suffice for a small brand.
Making the number useful
Report share of voice by topic, not just overall. A strong overall figure can hide a topic where you are invisible. Pair it with sources: if a competitor leads a topic, look at the cited pages. Add crawler logs to rule out access problems. Then use visitor analytics to check that visibility turns into visits.
No tool can guarantee a share of voice target or a recommendation. The number tells you direction. What you do after reading it is what changes it.