Who is ChatGPT recommending to your customers instead of you?

And right now, those answers name your competitors. In two business days, you get the full list: every brand named instead of you across ChatGPT, Claude, Perplexity, and Gemini, the prompts where you lost, and the fixes most likely to put you back in the answer. 30 to 50 buyer-intent queries, one flat-fee report.

Here is what lands in your inbox.

Prompt	ChatGPT	Claude	Perplexity	Gemini
best note-taking app for researchers	Brand A	Brand B	Brand A	Brand C
alternatives to Notion for small teams	Brand D	Brand A	Brand D	Brand E
tools for managing customer feedback	Brand F	Brand G	–	Brand F
best help desk for small SaaS	Brand H	Brand H	Your brand	Brand I
habit tracker that respects privacy	Brand J	Brand K	Brand J	Brand L
self-hosted analytics for indie founders	Brand M	Brand N	Brand M	–
email tools for newsletter creators	Brand O	Brand P	Brand O	Brand Q
Shopify alternatives for digital products	Brand R	Brand S	Brand R	Brand T

Three things, in one document.

Two things sit underneath all three. The matrix is diagnosed per LLM, because each one biases differently: ChatGPT cites vendor blogs, Perplexity skews household-name, Claude leans on parametric knowledge, Gemini on marketing tone. And every fix ships with a re-run protocol: which prompts to test again, in how many weeks, and the threshold for measuring lift.

From URL to report in 2 business days.

1. 01

   You submit a landing page URL.

2. 02

   I generate 30 to 50 buyer-intent prompts spanning discovery, comparison, and evaluation.

3. 03

   Every prompt runs across ChatGPT, Claude, Perplexity, and Gemini.

4. 04

   Mentions are tallied per brand per LLM into your visibility matrix.

5. 05

   Your site is inspected against the 2 to 3 most likely causes of low visibility, and fixes are ranked by effort.

What early runs surfaced.

Small sample from beta. The numbers will move; the patterns probably will not.

Four notes from the runs so far.

1. 1. LLMs return different brand sets for the same prompt. You cannot optimize for "AI search" as a single channel.
2. 2. Vendor-published comparison content ("X vs Y vs Z" listicles on the vendor's own blog) is currently the most reliable way in.
3. 3. API and consumer-chat outputs diverge. Consumer UIs may pull live web search; APIs often will not. Both need tracking.
4. 4. Fixes take weeks to surface in LLM outputs. Re-runs matter more than one-shots.

AEO.

Answer Engine Optimization. Getting picked up in AI answers.

GEO.

Generative Engine Optimization. Same idea, different acronym.

What it costs, plainly.

One audit, one fee. Want more runs? Buy them upfront and use them when you ship a redesign or shift positioning. No subscription, no auto-renew. Credits expire 12 months after purchase.

Add-on: Competitor benchmark (+€99). Add 2 to 3 named competitors to any audit and see exactly where they outrank you. Mention it in your email.

Single follow-up runs available ad-hoc at €39 each.

Who is running this.

I am Andrei. I am building Periscope after running early audits across a handful of indie SaaS products and watching the same patterns show up. If you want to talk before signing up, my inbox is open at periscope@surdu.eu.