Learn

Who decides which prompts AI visibility tools measure?

The customer and the tool decide. The people using AI engines don't. The engines don't publish what users type, so every tool works from a stand-in list: prompts the customer writes, prompts an AI generates from the customer's website, prompts rewritten from search keywords, or prompts bought from browser panels.

Of 91 tools with a tracking product, 22 don't say where their prompts come from at all. Early studies show the choice matters: list-style prompts lifted brand visibility by about 20% and keyword-style prompts by up to 25% in Peec AI's 2026 paper.

No published study has run the same brand measurement on all of these sources and compared the brands that get named. That is the gap our first research edition fills.

Who is speaking: Salience Index, paid by readers. Peec AI, which sells AI visibility tracking. Label: Verified for the 22 of 91, read on vendor pages. Verified for the Peec figures, read in the paper. Source: 91 vendor pages, captured 13 September 2026. Peec AI, SSRN, June 2026.

ChatGPT has a dozen indexes and five live experiments. Your prompt list should have no edits

The engine under your AI visibility number is being rebuilt in public. The only part of the measurement you control is the list of questions you ask it, and most of the advice on writing that list comes from the companies selling the tracking. Method · 21 September 2026 On 4

Cloudflare spent a year trying to split Googlebot. It settled for a line in robots.txt

On 15 September the company retired its one-click AI-blocking switch. What replaced it lets Google, Apple and Microsoft keep crawling, on a promise. Every other company's training bots stay locked out, and one name is missing from the announcement altogether. Engines · 16 September 2026 Somewhere in

Only 4 of 94 AI visibility tools mention Hindi. None say where their India prompts come from

We read the public pages of 94 AI visibility tools on 13 September 2026. Four name Hindi or Hinglish anywhere on those pages. Ten have an Indian address, rupee pricing or Indian clients. Not one, Indian or global, says on the pages we read how the prompts it runs for

Profound Raises $180 Million Series D

Bloomberg reported this morning that Profound, the New York company that tracks how brands show up in AI answers, has closed a $180 million Series D. The round values the company at $1.8 billion post-money and was led by Sequoia Capital and Kleiner Perkins. As always, every number

Read more

ChatGPT has a dozen indexes and five live experiments. Your prompt list should have no edits

Cloudflare spent a year trying to split Googlebot. It settled for a line in robots.txt

Only 4 of 94 AI visibility tools mention Hindi. None say where their India prompts come from

Profound Raises $180 Million Series D