What is the AI visibility tool that shows the actual AI answers, not just scores?

Which companies did the engines name as an AI visibility tool?

New Map's measurement counted 5 answers to this exact question: 2 from Gemini, 2 from Perplexity and 1 from ChatGPT. Profound was named in 3 of the 5. Otterly, Peec AI, SE Ranking and Semrush² were each named in 2, and Ahrefs in 1. New Map was named in 0.

Five answers is a first look, not a ranking. The 146-answer reading in the last column covers 68 different questions about AI visibility tools, not this one (New Map measurement).

The chart shows the same naming counts.

What sources do the engines cite in their actual answers?

New Map's measurement covered 146 answers to 68 questions about AI visibility tools (how New Map measures). That is one week's reading of 68 questions about AI visibility tools on 3 engines (146 answers), not a trend: Perplexity cited a source in only 7 of its 68 answers, and Claude Code's 10 answers are too few to read alone.

Gemini cited 9.3 sources per answer across its 68 answers and Perplexity 0.3 across its 68. Claude Code cited 7.3 across 10 (New Map measurement). Perplexity's low figure reflects that it cited nothing in most answers, so it is not a typical rate for that engine.

The same reading found reddit.com cited in 20 of 146 answers and youtube.com in 16 (New Map measurement). On the exact question, youtube.com was cited in 2 of 5 answers, and seven other domains in 1 each, among them ahrefs.com and medium.com.

Omnia data reported by Kevin Indig in May 2026 found 2.37% of 3.7M cited URLs in all of ChatGPT, Perplexity and AI Overviews for the same prompt. A single score cannot show which pages each engine leaned on.

The chart shows sources per answer by engine.

Why does a score alone hide what the answers say?

A score compresses answers that move. SparkToro found in January 2026, across 2,961 runs, that ChatGPT and Google AI repeated the same brand list in under 1 in 100 runs. Peec AI reported in September 2026, across 120M chats, that weekly moves under about 5 points can be chance.

Place changes the answer too. Chen and Koudas³ (University of Toronto, 16 October 2026, a preprint) audited 300 questions and found that ChatGPT follows the location an application declares, while Gemini follows the network's country. The businesses recommended changed with the sources.

Being cited is also not being recommended. Ahrefs found in July 2026 that 43% of 9,886 answers citing a brand's own event page never named the brand. A number that says "cited" cannot tell you which case you are in. Reading the answer can.

For the method behind such readings, see how a brand's visibility in AI answers is measured.

What should a tool show you before you trust its numbers?

Ask for the answer text, the sources behind it and the date it was drawn. Then check whether the numbers are exact or estimates.

Semrush noted in October 2026 that Google Search Console's Generative AI report shows impressions only, with no prompt-to-click attribution. Bing's Topics Portfolio, announced 13 October 2026, labels its topic sizes and citation volumes as approximate, directional signals.

On sources, the same Chen and Koudas preprint found, across 158,099 answers, that retrieval rank explained 32% of which pages ChatGPT cited within a site, while all page features together explained 2.8%. Our read: sources are worth showing. That study did not test any tool.

A tool worth trusting offers:

• Raw answer text per engine and per question
• Cited sources listed by origin
• Repeat runs, so noise is visible
• Location and language stated

Where does New Map fit when you want to read the actual answers?

New Map was named in 0 of the 5 answers to this question in its measurement, so no engine recommended it here. It measures what AI engines answer about a brand, drafts the fixes and places the brand in the press and creator sources the engines cite.

An AI account manager reads every answer and proposes the monthly plan. A human network places the brand with press, creators and communities. The facts we hold do not say whether the dashboard shows full answer text. Details are on plans and pricing.

The flow below shows the order of work, from question to placement.

What to do

1. Run your five most important buyer questions through each tool you shortlist and ask it to show the full answer text and cited sources for ChatGPT, Gemini and Perplexity, with dates.
2. Ask each tool for repeat runs of one question, since SparkToro's January 2026 study of 2,961 runs found brand lists rarely repeat.
3. Each week, ask the same 5 questions on each engine, save the answer text, and tally which brands and domains appear, as in how New Map measures.

Frequently asked

Sources

1. New Map measurement, How we measure, 19 October 2026
2. Semrush, How to build an AI visibility dashboard with Semrush, semrush.com, 12 October 2026
3. Chen and Koudas (University of Toronto), Location, Location, Generation: Where You Ask Matters for Generative Engine Optimization, arxiv.org, 16 October 2026
4. Microsoft Bing, Introducing Topics Portfolio (Beta) in AI Performance, blogs.bing.com, 13 October 2026
5. Chen et al. (arXiv), What Drives AI Citations? Content, Retrieval and Model Knowledge in Generative Engine Optimization (GEO), arxiv.org, 10 October 2026