About AmppScaling

What AmppScaling does

AmppScaling is a generative engine optimization (GEO) consultancy for companies whose buyers ask AI assistants for recommendations. We measure how often those assistants name a company, compare that against the competitors being named instead, and then do the work to close the gap.

Why the company exists

Buyers now shortlist vendors by asking an AI assistant before they visit a website or fill out a form. The assistant names three to five companies. Companies absent from those answers lose deals they never see, because there is no lost RFP to point at. Ranking well on Google does not mean being cited by ChatGPT.

How we measure

We identify the questions a company's buyers actually ask, then run each question ten times across three engines — three runs on Perplexity, three on OpenAI, four on Gemini. Each answer is classified as Cited (you are recommended), Mentioned (named but not chosen), or Absent. The result is a 0-100 AI Visibility Score, shown beside the same score for the competitor being recommended instead of you. The score is simply how many of those ten runs named you, so the denominator is never in doubt.

Why every run is grounded

Every run is grounded, meaning the engine performs a live search before it answers. An ungrounded model answers from memory, which measures whether it absorbed you during training rather than whether it recommends you today. Those are different questions, and only one of them loses you deals.

Why your competitor is measured on the same runs

Your competitor is scored on the identical runs — same questions, same engines, same day. Comparing a number from one set of runs against a number from another set is the most common way these reports mislead, and it is not something we do. Appearances that only cite your own website are reported separately, never folded into the headline number. An engine quoting your homepage back at you is not a recommendation.

What makes this different

Most AI visibility claims come from a single screenshot of a single answer. AI answers vary between runs, so one check measures nothing. Every AmppScaling number is sampled at least thirty times and shipped with the underlying transcripts. Any AI visibility figure has a shelf life of about two weeks. The sources these engines cite turn over fast, so a baseline is re-measured before it is used to make a decision rather than quoted from a slide.

Who runs AmppScaling

AmppScaling was founded in 2026 by Gabriel Cervantes. Engagements are delivered directly by the founder, not routed to an account manager.

Company facts