Tony Uphoff ran a test this month that every SEO and brand manager should copy before next year’s planning starts. He asked ChatGPT, Gemini, Perplexity, and Google AI Overviews the questions his prospective clients ask, then counted how often the answer engines recommended the two businesses he launched in February.
The results were humbling. Across 59 answers, his newsletter, Uphoff on Media, showed up in zero of 12 category answers. His firm, Uphoff Advisory, showed up in one of 43. Yet every engine that answered “Who is Tony Uphoff?” knew his career.
Tony said “The models know who I am. They almost never recommend what I’ve built.” That gap, between being known and being recommended, is the measurement problem I think most of our industry is still getting wrong. Tony’s test shows you how to find it in your own brand, and three other posts he published in October explain why your CFO, your next acquirer, and your next employer are about to care about it.
From A Certain Point Of View, We’re On The Same Beat
A disclosure first. Tony and I worked together at Ziff-Davis from 1988 to 1993, and he’s since been CEO of five businesses. He was brought in to transform Business.com and ThomasNet.com, led the carve-out of Pipeline360, where he now serves as a board advisor, and today writes Uphoff on Media while running Uphoff Advisory.
His audience is B2B media, marketing, and technology operators, not SEO specialists. But read his last four posts back to back, and you’ll find a five-time CEO working through the same questions this column keeps asking about AI search, from the other side of the table.
Bill Ziff taught a generation of us at Ziff-Davis to separate events from trends. Tony’s October 7 headline, “AI Stopped Being a Story. Now It’s in Your Multiple,” is that lesson in 10 words. When AI shows up in the valuation of a company, it has stopped being an event.
See also: The Click Economy Is Over: How AI Search Is Forcing Publishers To Rethink Revenue
Search Changed Jobs, And The Scoreboard Didn’t
Tony’s argument starts with a sequence change in B2B buying. “The machine drafts the shortlist. The human validates it.”
The survey data backs him up. In G2’s March 2026 survey of 1,076 B2B software buyers, 51% said they now start their research with an AI chatbot more often than with Google, up from 29% in G2’s 2025 report. 69% said a chatbot led them to choose a different vendor than they had planned, and one in three bought from a vendor they had never heard of before.
Here is where Tony makes the point I’d underline for every SEO reading this.
“For two decades the scoreboard was discoverability. SEO taught every marketing team to rank, earn the click and get found. Most companies carried that scoreboard straight into AI.”
They ask a model about themselves by name, get an accurate answer, and conclude they’re covered. Tony calls that a false positive, and his own results prove it. A well-known founder or a legacy brand makes the business feel visible. Ask what a buyer would actually ask, and the company rarely makes the list.
In my columns on YouGov’s AI brand research, I’ve kept coming back to one idea. Awareness is not the metric that tells you who is winning. Consideration is. Tony’s discoverable-versus-recommended split is the same funnel distinction, measured inside the answer engine.
If brand recognition is doing that little of the work, then a report built on branded queries is measuring the wrong thing.
See also: Brand Protection In AI Search: How To Audit And Defend Your Brand’s Identity
Answer Engines Describe The Company You Used To Be
The second finding should worry anyone who has been through a rebrand, a merger, or a leadership change. Tony called out that “all three engines that described me still call me CEO of Pipeline360.” Only one mentioned the firm he runs now. He explained, “The sources included org-chart aggregators, executive databases and old press releases.”
Tony’s rule of thumb is blunt. Answer engines default to what the web has said about you most often, “and most often usually means longest ago.”
That stale record costs twice, because the machine uses it to decide whether to recommend you and the human uses it to verify the machine. Few teams are watching. A Corporate Ink study reported by eMarketer in May found only 29% of marketers actively track and measure AI visibility.
There’s a second lesson buried in Tony’s test. When he asked the engines “What is Demand-as-a-Service?” all three defined it accurately and cited Pipeline360’s own pages. None mentioned Tony, who helped create the strategy. He’s fine with that, because the company owns the category language. But his conclusion is the one content strategists should tape to the monitor. “The models credit whoever publishes the clearest explanation, not whoever had the idea first.”
See also: AI Visibility Measurement: What To Track & What To Ignore
AI Shortlists Are Built From Other People’s Shortlists
The one time Uphoff Advisory made a list, ChatGPT ranked it second of four for “Who advises founders of B2B information and data businesses?” The broad questions returned nothing.
The citations explain why. On category questions, the models rarely cited the firms they named. They cited roundups, the “best AI newsletters” and “best B2B marketing podcasts” lists published by Zapier, Content Allies, and others.
Tony points to a May 2026 study by Maryanna Franco and João da Silva on Friction AI’s platform that found the same mechanism at scale across 14,140 queries about activewear brands. Nike and New Balance share the same Google Knowledge Graph description, yet Nike appeared in about 71% of athleisure recommendation answers and New Balance almost never did.
This is where I’d push back on Tony a little. Nineteen prompts run for one week is a snapshot, and AI answers shift from run to run. One appearance in 43 could be a floor or it could be noise, and a single week can’t tell you which. His plan to rerun the same prompts in January and publish the numbers either way is exactly right. I’d go further and track it weekly, because Citation Share of Voice only means something as a trend line.
Renting Your Customer Relationship From Search Is Now A Discount
Tony’s October 7 post introduces the Q3 2026 Uphoff Value Index, his quarterly benchmark of what buyers pay for B2B information, marketing, and technology businesses. This quarter he added a score for AI Disruption Protection built on four questions. The third one stopped me.
“Relationship ownership. Do you own the customer relationship, or rent it from search and platforms?”
Tony cites Collingwood’s 2026-27 Market Report, a survey of about 100 active acquirers, in which 91% called the AI resilience of a business model essential or important and 74% said the same of a direct relationship with the customer. In his worked example, two data businesses with identical P&Ls are valued at 5.6 times EBITDA ($28 million) and 4.0 times ($20 million). The only difference is whether buyers believe AI can route around them.
I read that third question as the most direct challenge to how many of us have pitched organic search for 20 years. To an acquirer, traffic that arrives from Google and leaves without a login, a subscription, or a first-party record is a rented relationship. In my view, the SEO teams that thrive in this cycle will be the ones that can show how search converts rented visits into owned relationships, not the ones with the biggest session counts.
Revenue Per Employee Is The Scoreboard Your CFO Is Reading
Tony’s October 19 post, “Before You Flatten the Org Chart, Make Sure You Can Measure What AI Delivers,” uses Block as its case study. By his math from Block’s reported figures, the company produced $10.36 billion of gross profit with 10,205 people in 2025, about $1 million per person. Its raised 2026 guidance of $12.51 billion with just under 6,000 people works out to about $2.1 million per person.
The benchmarks he cites matter more for most Search Engine Journal readers. The average marketing agency does about $163,000 of revenue per employee. The median public software company does about $395,000. In Tony’s index, two agencies with identical P&Ls, one with 68 people and one with 40, are now about $1 million apart in value.
Tony describes a spectrum from honest signal to performative signal. Contract value, expansion revenue, and churn sit near money. Form fills and self-reported data sit at the cheap-to-fake end. Then this.
“A lead score isn’t a fact about someone’s business. It’s a fact about how well your content ranks.”
That’s an uncomfortable read for anyone whose monthly report leads with rankings and organic sessions. If you want to move from SEO manager to director, my advice is to rebuild your report, so it starts at the honest end of Tony’s spectrum and works backward to the search metrics that explain it.
Give Your GEO Pilot A Job
Tony’s Oct. 14 post, “Your AI Pilot Doesn’t Have a Job,” borrows a phrase from Darshan Deshmukh of ProcureAbility on the Art of Procurement podcast, “death by a thousand pilots.” The diagnosis is that teams lead with the technology instead of the outcome. Tony cites BCG’s 2024 finding that roughly 70% of the challenges in getting value from AI come from people and process, 20% from technology, and only 10% from algorithms.
SEO teams are buying AI visibility trackers right now, often on a corporate card, which is precisely the pattern Tony warns about. His Application Test has three gates, cleared before any money moves. State the job in plain words with no vendor name in it. Name the owner and the baseline. Set the one signal you’ll check in 90 days, and decide in advance whether you’ll scale, fix, or stop.
“Track our AI mentions” fails the first gate. “Get us onto the AI shortlist when buyers ask for our narrowest category phrase” passes it, and Tony’s own 1-in-43 is what a baseline looks like.
How To Apply This To Your SEO Strategy
Tony closes with a five-move sequence. Here’s how I’d adapt it for SEO teams:
Add a recommendation rate to your scorecard, and report it first. Write 10 to 15 prompts the way your buyers ask them, built around the category, the problem, and “who should I hire for.” Add two or three branded prompts. Run them across ChatGPT, Gemini, Perplexity, and AI Overviews, and track how often you make the list and whether the description is accurate. Run it weekly, not once, so you’re reading a trend rather than a snapshot.
Fix the record the models read before you chase the shortlist. Pull the sources the engines cite when they describe your brand. Update the pages you control, including bios, LinkedIn, and the press page, and request corrections from the aggregators and directories you don’t. After any rebrand or leadership change, assume the models haven’t heard.
Own the narrow phrase, then earn a place on the lists. Pick the most specific way a buyer describes your problem and publish the canonical explanation under your brand. Then find the 10 to 20 roundup and comparison pages the engines cite in your category and give someone a target for getting on them. That’s a digital PR job, and it’s the job SEO and PR should have been sharing all along.
Translate every AI visibility gain into an owned relationship. Tie recommendation rate to signups, demo requests, subscribers, and pipeline, so your CFO sees the honest end of the spectrum rather than another traffic chart.
Same Problem, Different Galaxy
Tony and I haven’t worked at the same company or attended the same meeting for 33 years. But, from a certain point of view, we never stopped working on the same problem, which is getting the right name in front of the buyer before the buyer goes looking.
What’s changed is who builds the shortlist. My bet is that the teams still reporting rankings first will spend 2027 explaining a pipeline gap that their dashboards never showed them. The models are drafting your buyer’s list right now. Make sure you’re measuring whether you’re on it.
More Resources:
- Decision Coverage: Why AI Recommends Some Brands And Not Others
- 3 Steps To Win B2B Brand Consideration, Day Zero Decides The Shortlist
- How AI Agents Decide Which Brands To Recommend: Trust Is The New Ranking Factor
Featured Image: patpitchaya/Shutterstock
