AI SEO Tools Benchmark, edition 1
Promptwatch in the AI SEO Tools Benchmark
Tracks brand mentions in answers from ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude and other AI engines, with daily checks. Not for data exports on the entry plan; exports start at the Professional plan.
Promptwatch, edition 1
- Category
- AI visibility tracking
- Status
- Live
- Website
- promptwatch.io
- Measured
- 1,199 answers, collected on 16 October 2026 (India time); one of them a ChatGPT test answer from 13 October
- Asked from
- United States, in English
- Accuracy
- Listed, not yet accuracy-tested
- Written by
- Chandan Bakshi, Founder, SEO Informatica Labs
How often AI answers named Promptwatch
We asked the same 25 buyer questions on four surfaces. A rate counts the answers that named Promptwatch, out of every answer on that surface. It is a mention rate over these runs, not a rank in AI.
- ChatGPT
- 1%: named in 9 of 750 answersNamed in answers to 4 of 25 questions. Its own site cited in 7 of 750.
- Perplexity
- 0%: named in 0 of 400 answersNamed in answers to 0 of 25 questions. Its own site cited in 0 of 400.
- Google AI Overviews
- 4%: named in 1 of 24 overviewsNamed in answers to 1 of 25 questions. Its own site cited in 0 of 24.
- Google AI Mode
- 4%: named in 1 of 25 answersNamed in answers to 1 of 25 questions. Its own site cited in 0 of 25.
Price, plan and free option
The lowest published plan, as the vendor's own page showed it on the date below. Prices change; follow the source before you buy.
- Starting price
- $95 a monthEssential plan. Annual billing toggle; annual price not recorded. 1 project, 50 prompts, 6,000 responses a month, 4 models, daily, 500 agent credits
- Free option
- Free plan. Free Explore tier in docs: ChatGPT only, 10 prompts, 1 project; paid plans have a 7-day trial
- Verified
- Verified on 16 October 2026, at promptwatch.com/pricing
- How it was read
- Read twice: a second reader checked the same pages and confirmed it.
- Our reading of the vendor's prices and limits, on 16 October 2026
- Essential $95/month: 1 project, 50 prompts, 6,000 responses/month, 4 models, daily, 500 agent credits, MCP and API. Professional $245/month: 2 projects, 150 prompts, 18,000 responses. Business $579/month: 5 projects, 350 prompts, 42,000 responses. Monthly/annual toggle; agency and custom tracks separate.
Engines tracked
The AI engines the vendor says the product tracks, on at least one plan, as its own pages stated them.
- Engines
- ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Gemini, Claude, Microsoft Copilot, Grok, DeepSeek, Meta AI, Other
- Per plan, as we read the vendor's pages
- Engines are named. The pricing page lists ChatGPT, Perplexity, Google AI Overviews, AI Mode, Claude, Gemini, Meta Llama, DeepSeek, Grok, OpenCode and Claude Code, and the docs add Copilot, Alexa, Codex and Mistral. What each plan includes is contradictory. On the pricing page, each paid brand plan card (Essential, Professional, Business) says 'Track 4 models', with a tooltip naming ChatGPT, Perplexity, Gemini and Claude, while agency and enterprise say 'Track all models'. Yet the same page's comparison table ticks all 11 models for every paid plan. The docs (Choosing Your Models) say there is no model cap and every paid plan unlocks all active models, with Explore limited to ChatGPT. Source
Status and ownership
Live.
Free checker
Not run: it needs an account.
What it is for, and not for
From the vendor's own pages, read on 16 October 2026. A prompt is a question a tool asks AI engines for you; a citation is a link to a site among an answer's sources; an API is a way for your own software to pull a tool's data.
- For
- Tracks brand mentions in answers from ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude and other AI engines, with daily checks.
- Not for
- Not for data exports on the entry plan; exports start at the Professional plan.
Scorecard
The openness score rates what a vendor publishes about its product (engines, method, price, data freshness, exports), not how good the product is.
- Openness score
- 50.3 of 60
- Accuracy and setup time
- Accuracy (30) and setup time (10): pending access. Listed, not yet accuracy-tested.
- Coverage
- 9 of 15: which AI engines it tracks, and on which plans
- Transparency
- 11.3 of 15: disclosed run counts and sampling, and whether the models and the surface (app or API) are named
- Price
- 15 of 15: a published starting price with plan limits, and a free tier
- Freshness
- 10 of 10: how often the data updates
- Exports
- 5 of 5: data exports and a documented API
Every fact behind the score, with its page
- Starting price and plan limits: level 1
- Essential $95/month: 1 project, 50 prompts, 6,000 responses/month, 4 models, daily, 500 agent credits, MCP and API. Professional $245/month: 2 projects, 150 prompts, 18,000 responses. Business $579/month: 5 projects, 350 prompts, 42,000 responses. Monthly/annual toggle; agency and custom tracks separate. SourceConfirmed by a second reading.
- Engines tracked: level 0.6
- Engines are named. The pricing page lists ChatGPT, Perplexity, Google AI Overviews, AI Mode, Claude, Gemini, Meta Llama, DeepSeek, Grok, OpenCode and Claude Code, and the docs add Copilot, Alexa, Codex and Mistral. What each plan includes is contradictory. On the pricing page, each paid brand plan card (Essential, Professional, Business) says 'Track 4 models', with a tooltip naming ChatGPT, Perplexity, Gemini and Claude, while agency and enterprise say 'Track all models'. Yet the same page's comparison table ticks all 11 models for every paid plan. The docs (Choosing Your Models) say there is no model cap and every paid plan unlocks all active models, with Explore limited to ChatGPT. SourceThe first two readings differed; a third reading settled it.
- Disclosed run counts and sampling: level 1
- On each run day every prompt is paired with every selected model and each pair yields one response (models x prompts x runs = responses), daily by default. Visibility Score averages per-response 0-100 scores across all analyzed responses, absent responses counted as zero; Share of Voice is mentions over total mentions. SourceConfirmed by a second reading.
- Model and surface disclosed: level 0.5
- Live-search platforms collected from the real product (automated browser sessions on ChatGPT; live Gemini, Perplexity, Copilot, Alexa; real Google results pages for AI Overviews and AI Mode); Claude, DeepSeek, Mistral, Grok via provider APIs. Tracked model versions are not stated (API reference only shows example IDs). SourceConfirmed by a second reading.
- Data freshness: level 1
- Daily by default on every plan and trials; slower cadences only if sales enables a custom frequency feature; new setups are fast-tracked the same day. SourceConfirmed by a second reading.
- Exports and API: level 1
- Responses and 20+ other tables export to Excel or CSV (10,000-row cap on responses) and PDF reports, Professional and above; Data Studio integration; MCP and a public API reference (v1 and v2, OpenAPI) at promptwatch.com/docs/api-reference, API on all paid plans. SourceConfirmed by a second reading.
- Free tier: level 1
- The docs (Plans and billing, Choosing your models) describe Explore as the free entry tier: ChatGPT only, no Agent Chat, no site audits. The pricing page's own structured data lists an Explore offer at $0 per month, and its page code sets Explore at 1 project, 10 prompts and 310 responses per month, tracked daily. The Quickstart says new users choose a plan and that paid plans include a trial. The data-freshness doc says monitors keep running for organisations that have never subscribed. No time limit is stated. The visible pricing table shows only paid plans with a 7-day free trial (card terms not stated). The free AI Visibility Report needs no account, but it currently asks for an email (page setting requireEmail is true), so it does not count. Browser run, 16 October 2026: the free AI Visibility Report opened an account form (name, work email, terms) before any result, so it is not a no-login checker; the level rests on the free Explore tier in the docs. SourceThe first two readings differed; a third reading settled it.
Each fact gets a level from 0 (not published) to 1 (published in full), and each part of the score is its points times those levels. How the levels and points work.
Read on 16 October 2026.
The Clarity Audit
Want to know who AI names in your market?
The Clarity Audit is free. We look at your website and your market, show which businesses Google and AI answers name for the searches your buyers make, and list the three fixes to consider first. No call needed.