supereo Index · method
How the Index is counted
The supereo Index reports what AI engines recommend when buyers ask about software. It doesn't test or rate the tools: a brand's share says how often the engines recommended it, not how good it is.
The questions
Each category is a set of buyer scenarios: a buyer, the job they need done and the words they search for. Each scenario is asked several ways:
- Problem first: describes the problem and names no tool.
- Best tool: asks for the best tool for the job.
- Comparison: weighs named tools or an alternative to one.
- How to: asks how to get the job done.
- Names an AI client: says which AI client the asker uses, such as Claude Code or ChatGPT.
- Asks for AI: asks how to do it with AI, naming no client.
- No AI: never mentions AI.
| Category | Scenarios | Questions | Search keywords | Answers | Run |
|---|---|---|---|---|---|
| Agent toolkits and MCP servers | 10 | 46 | 20 | 250 | 19 Oct 2026 |
| AI lead generation and SDRs | 7 | 33 | 14 | 178 | 19 Oct 2026 |
| Intent data | 8 | 39 | 11 | 206 | 19 Oct 2026 |
| AI email assistants | 8 | 39 | 16 | 210 | 19 Oct 2026 |
| AI visibility tools | 8 | 39 | 16 | 211 | 19 Oct 2026 |
The engines
Every question goes once to each of ChatGPT, Gemini, Google AI Mode, Perplexity and Claude, on the run date. Google AI Overviews answer searches rather than questions, so they were asked each scenario's search keywords instead. Each engine is asked the way a buyer would reach it:
- ChatGPT: its app, signed out.
- Gemini: its app, signed out.
- Google AI Mode: Google's live results.
- Perplexity: its API, with web search on and set to the US.
- Claude: its API, with web search on and set to the US.
- Google AI Overviews: Google's live results, for the scenarios' search keywords.
What gets counted
- Recommended
- A grader reads every answer and lists the brands it recommends, in order, and the pages it cites. Brand names are the ones the grader records, with obvious variants merged.
- Share
- Share of an engine's answers in this category that recommend the brand; range is the 95% Wilson interval.
- All chat engines
- The answers from ChatGPT, Gemini, Google AI Mode, Perplexity and Claude, pooled. Google AI Overviews stay out of the pool and get a table of their own.
- Average place
- A brand's mean position among the brands an answer recommends, 1 being first.
- Cited pages
- The pages an engine's answers cite, each with how many answers cite it.
- Left out
- Questions that named one of our products, and a few that named one of our data suppliers; AI assistants and platforms (Gmail, Slack, LinkedIn and similar) are not counted as recommendations.
Who publishes it
The supereo Index is published by supereo, part of superagnt_. Our own products in each category:
- Agent toolkits and MCP servers: superagnt_.
- AI lead generation and SDRs: superagnt_ and supergtm.
- Intent data: superradar.
- AI email assistants: inboxagnt.
- AI visibility tools: supereo.
We count our own products on the same answers as every other brand, and each category page says in plain words which of ours is in it and how often the engines named or cited it.
Limits
- One run: each question was asked once on each engine. Asked again, an engine may answer differently.
- Small numbers: an engine answers a few dozen questions per category, so a share is a rough figure. Read the range as well as the share.
- Each list stops at the brands recommended most. A brand tied with the last one listed may be left off.