← All Frameworks

Discoverable Is Not Recommended

Two tests: does AI describe you accurately, and does it put you on the shortlist when a buyer asks for your category? Find your square and fix what it calls for.

Two-by-two matrix plotting the discoverability test, whether AI describes a company accurately, against the recommendation test, whether AI puts it on the buyer's shortlist, with squares Invisible, The Borrowed Shelf, Discoverable Not Recommended, and The Default Pick

How do I know if AI is recommending my company to B2B buyers?

Discoverable Is Not Recommended is a two-test check of how AI treats your company. Test one: ask a model about you by name and see whether the answer is right and current. Test two: ask for your category the way a buyer would and see whether you make the list. Together they place you in one of four squares.

Most B2B companies carried the discoverability scoreboard (rankings, traffic, mentions) straight into AI. A model that describes you accurately but never shortlists you is a false positive. The brand feels visible, so nobody measures the list it’s missing from.

How to use it

  1. Write 10 to 15 prompts the way your buyer asks them.

    The category, the problem, “who should I hire for.” Add two or three that ask about you by name.

  2. Run them across the major models.

    ChatGPT, Gemini, Perplexity and Google AI Overviews, for yourself and your top three competitors.

  3. Track two numbers.

    How often you make the list, and whether the description is right. Report the list rate first, ahead of traffic.

  4. Find your square and start with what it calls for.

    Invisible: fix the record the models read. The Borrowed Shelf: publish the definitive explanation of your ideas under your name. Discoverable, Not Recommended: own the most specific question a buyer asks, then earn a place on the roundups the models cite. The Default Pick: rerun both tests every quarter.

Questions

What is the difference between discoverable and recommended?

Discoverable means a model describes you accurately when asked by name. Recommended means it puts you on the shortlist when a buyer asks for your category without naming you.

How often should I run the tests?

Monthly to start. Quarterly once you are on the list, and again after any leadership change, rebrand or new offer.

Related: The Signal-Honesty Spectrum · The Agent Exposure Test · The Windshield Test

Discoverable Is Not Recommended is an Uphoff Advisory Framework created by Tony Uphoff, five-time B2B CEO and founder of Uphoff Advisory, first published in Uphoff on Media on October 21, 2026. Read the full piece: I Asked AI Who to Hire. It Recommended Me. Once

Framework Library

Download framework

By downloading, you agree to be contacted by Uphoff Advisory about this framework.