Uphoff ran a 59-answer AI test. The shortlist named him once: Reveneau

Why it mattersMany teams measure AI visibility by asking the model about their own brand. The model can pass that test and still never recommend the brand to a buyer.

Ask an AI engine who you are and it has your bio right. Ask who it would recommend for your category and your name almost never comes up. That is the gap Tony Uphoff tried to measure. Greg Jarboe wrote up the result in Search Engine Journal.

Uphoff is a five-time CEO. He runs a newsletter, Uphoff on Media, and an advisory firm, Uphoff Advisory. He ran 19 prompts for one week on four AI engines: ChatGPT, Gemini, Perplexity and Google AI Overviews. That is 59 answers in total.

Jarboe reports the shortlist numbers. Uphoff on Media appeared in zero of 12 category answers. Uphoff Advisory appeared in one of 43 category answers. Every engine that answered "Who is Tony Uphoff?" knew his career.

Known, but rarely recommended

Jarboe quotes Uphoff: "The models know who I am. They almost never recommend what I have built." Uphoff Advisory showed up once. ChatGPT ranked it second of four. The question was "Who advises founders of B2B information and data businesses?" Broad questions returned nothing.

Jarboe notes who the models did cite on category questions. They cited roundup lists. "Best AI newsletters" and "best B2B marketing podcasts" lists published by Zapier, Content Allies and others. The named firms were rarely cited for themselves.

Stale sources decide what gets said

A second finding in Jarboe's write-up. All three engines that described Uphoff still called him CEO of Pipeline360, the company he carved out and left. Only one mentioned the firm he runs now. Uphoff told Jarboe the sources included org-chart aggregators, executive databases and old press releases. His rule of thumb: "Answer engines default to what the web has said about you most often, and most often usually means longest ago."

Jarboe adds a caution. Nineteen prompts over one week is a snapshot. AI answers shift from run to run. He writes that one appearance in 43 "could be a floor or it could be noise". He backs Uphoff's plan to rerun the same prompts in January.

Who is reading the model

Why buyers now ask: Jarboe cites a G2 survey of 1,076 B2B software buyers. G2 reports 51% start research with an AI chatbot more often than with Google. That is up from 29% in G2's prior report. G2 reports 69% changed vendor based on a chatbot's answer. One in three bought from a vendor they had not heard of before.

Why so few teams measure it. Jarboe cites a Corporate Ink study reported by eMarketer. The study found 29% of marketers actively track and measure AI visibility.

Jarboe also points to a Franco and da Silva study on Friction AI's platform. It ran 14,140 queries about activewear brands. The study reports Nike appeared in about 71% of athleisure recommendation answers. New Balance almost never did. The two share the same Google Knowledge Graph description.

A brand audit that asks the model about itself can pass without a single recommendation to a buyer. The test to run is Uphoff's. Write the questions a buyer in your category actually asks. Run them on every engine your market uses. Count how often the firm appears, and how often it is cited by name.

Source

AI Knows Your Brand, It Rarely Puts You On The Shortlist, Greg Jarboe in Search Engine Journal.

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.