Tony Uphoff ran a check this month that each search engine optimization and model supervisor ought to copy earlier than subsequent 12 months’s planning begins. He requested ChatGPT, Gemini, Perplexity, and Google AI Overviews the questions his potential shoppers ask, then counted how typically the reply engines really helpful the 2 companies he launched in February.
The outcomes have been humbling. Throughout 59 solutions, his e-newsletter, Uphoff on Media, confirmed up in zero of 12 class solutions. His agency, Uphoff Advisory, confirmed up in certainly one of 43. But each engine that answered “Who’s Tony Uphoff?” knew his profession.
Tony mentioned “The fashions know who I’m. They nearly by no means advocate what I’ve constructed.” That hole, between being recognized and being really helpful, is the measurement drawback I believe most of our business continues to be getting mistaken. Tony’s check reveals you discover it in your individual model, and three different posts he printed in October clarify why your CFO, your subsequent acquirer, and your subsequent employer are about to care about it.
From A Sure Level Of View, We’re On The Identical Beat
A disclosure first. Tony and I labored collectively at Ziff-Davis from 1988 to 1993, and he’s since been CEO of 5 companies. He was introduced in to rework Enterprise.com and ThomasNet.com, led the carve-out of Pipeline360, the place he now serves as a board advisor, and right this moment writes Uphoff on Media whereas operating Uphoff Advisory.
His viewers is B2B media, advertising, and know-how operators, not search engine optimization specialists. However learn his final 4 posts again to again, and also you’ll discover a five-time CEO working by the identical questions this column retains asking about AI search, from the opposite facet of the desk.
Invoice Ziff taught a era of us at Ziff-Davis to separate occasions from tendencies. Tony’s October 7 headline, “AI Stopped Being a Story. Now It’s in Your A number of,” is that lesson in 10 phrases. When AI reveals up within the valuation of an organization, it has stopped being an occasion.
See additionally: The Click on Financial system Is Over: How AI Search Is Forcing Publishers To Rethink Income
Search Modified Jobs, And The Scoreboard Didn’t
Tony’s argument begins with a sequence change in B2B shopping for. “The machine drafts the shortlist. The human validates it.”
The survey knowledge backs him up. In G2’s March 2026 survey of 1,076 B2B software program patrons, 51% mentioned they now begin their analysis with an AI chatbot extra typically than with Google, up from 29% in G2’s 2025 report. 69% mentioned a chatbot led them to decide on a unique vendor than they’d deliberate, and one in three purchased from a vendor they’d by no means heard of earlier than.
Right here is the place Tony makes the purpose I’d underline for each search engine optimization studying this.
“For 20 years the scoreboard was discoverability. search engine optimization taught each advertising crew to rank, earn the clicking and get discovered. Most corporations carried that scoreboard straight into AI.”
They ask a mannequin about themselves by identify, get an correct reply, and conclude they’re lined. Tony calls {that a} false optimistic, and his personal outcomes show it. A well known founder or a legacy model makes the enterprise really feel seen. Ask what a purchaser would truly ask, and the corporate not often makes the checklist.
In my columns on YouGov’s AI model analysis, I’ve saved coming again to 1 thought. Consciousness isn’t the metric that tells you who’s profitable. Consideration is. Tony’s discoverable-versus-recommended cut up is similar funnel distinction, measured inside the reply engine.
If model recognition is doing that little of the work, then a report constructed on branded queries is measuring the mistaken factor.
See additionally: Model Safety In AI Search: How To Audit And Defend Your Model’s Id
Reply Engines Describe The Firm You Used To Be
The second discovering ought to fear anybody who has been by a rebrand, a merger, or a management change. Tony referred to as out that “all three engines that described me nonetheless name me CEO of Pipeline360.” Just one talked about the agency he runs now. He defined, “The sources included org-chart aggregators, government databases and previous press releases.”
Tony’s rule of thumb is blunt. Reply engines default to what the online has mentioned about you most frequently, “and most frequently normally means longest in the past.”
That stale file prices twice, as a result of the machine makes use of it to determine whether or not to advocate you and the human makes use of it to confirm the machine. Few groups are watching. A Company Ink research reported by eMarketer in Might discovered solely 29% of entrepreneurs actively monitor and measure AI visibility.
There’s a second lesson buried in Tony’s check. When he requested the engines “What’s Demand-as-a-Service?” all three outlined it precisely and cited Pipeline360’s personal pages. None talked about Tony, who helped create the technique. He’s tremendous with that, as a result of the corporate owns the class language. However his conclusion is the one content material strategists ought to tape to the monitor. “The fashions credit score whoever publishes the clearest clarification, not whoever had the thought first.”
See additionally: AI Visibility Measurement: What To Observe & What To Ignore
AI Shortlists Are Constructed From Different Folks’s Shortlists
The one time Uphoff Advisory made an inventory, ChatGPT ranked it second of 4 for “Who advises founders of B2B data and knowledge companies?” The broad questions returned nothing.
The citations clarify why. On class questions, the fashions not often cited the companies they named. They cited roundups, the “finest AI newsletters” and “finest B2B advertising podcasts” lists printed by Zapier, Content material Allies, and others.
Tony factors to a Might 2026 research by Maryanna Franco and João da Silva on Friction AI’s platform that discovered the identical mechanism at scale throughout 14,140 queries about activewear manufacturers. Nike and New Steadiness share the identical Google Data Graph description, but Nike appeared in about 71% of athleisure suggestion solutions and New Steadiness nearly by no means did.
That is the place I’d push again on Tony a little bit. Nineteen prompts run for one week is a snapshot, and AI solutions shift from run to run. One look in 43 might be a ground or it might be noise, and a single week can’t inform you which. His plan to rerun the identical prompts in January and publish the numbers both method is precisely proper. I’d go additional and monitor it weekly, as a result of Quotation Share of Voice solely means one thing as a development line.
Renting Your Buyer Relationship From Search Is Now A Low cost
Tony’s October 7 put up introduces the Q3 2026 Uphoff Worth Index, his quarterly benchmark of what patrons pay for B2B data, advertising, and know-how companies. This quarter he added a rating for AI Disruption Safety constructed on 4 questions. The third one stopped me.
“Relationship possession. Do you personal the client relationship, or lease it from search and platforms?”
Tony cites Collingwood’s 2026-27 Market Report, a survey of about 100 lively acquirers, by which 91% referred to as the AI resilience of a enterprise mannequin important or necessary and 74% mentioned the identical of a direct relationship with the client. In his labored instance, two knowledge companies with equivalent P&Ls are valued at 5.6 instances EBITDA ($28 million) and 4.0 instances ($20 million). The one distinction is whether or not patrons imagine AI can route round them.
I learn that third query as probably the most direct problem to how many people have pitched natural seek for 20 years. To an acquirer, site visitors that arrives from Google and leaves and not using a login, a subscription, or a first-party file is a rented relationship. For my part, the search engine optimization groups that thrive on this cycle would be the ones that may present how search converts rented visits into owned relationships, not those with the most important session counts.
Income Per Worker Is The Scoreboard Your CFO Is Studying
Tony’s October 19 put up, “Earlier than You Flatten the Org Chart, Make Certain You Can Measure What AI Delivers,” makes use of Block as its case research. By his math from Block’s reported figures, the corporate produced $10.36 billion of gross revenue with 10,205 individuals in 2025, about $1 million per particular person. Its raised 2026 steering of $12.51 billion with just below 6,000 individuals works out to about $2.1 million per particular person.
The benchmarks he cites matter extra for many Search Engine Journal readers. The typical advertising company does about $163,000 of income per worker. The median public software program firm does about $395,000. In Tony’s index, two businesses with equivalent P&Ls, one with 68 individuals and one with 40, are actually about $1 million aside in worth.
Tony describes a spectrum from trustworthy sign to performative sign. Contract worth, growth income, and churn sit close to cash. Kind fills and self-reported knowledge sit on the cheap-to-fake finish. Then this.
“A lead rating isn’t a reality about somebody’s enterprise. It’s a reality about how effectively your content material ranks.”
That’s an uncomfortable learn for anybody whose month-to-month report leads with rankings and natural classes. If you wish to transfer from search engine optimization supervisor to director, my recommendation is to rebuild your report, so it begins on the trustworthy finish of Tony’s spectrum and works backward to the search metrics that specify it.
Give Your GEO Pilot A Job
Tony’s Oct. 14 put up, “Your AI Pilot Doesn’t Have a Job,” borrows a phrase from Darshan Deshmukh of ProcureAbility on the Artwork of Procurement podcast, “demise by a thousand pilots.” The prognosis is that groups lead with the know-how as a substitute of the end result. Tony cites BCG’s 2024 discovering that roughly 70% of the challenges in getting worth from AI come from individuals and course of, 20% from know-how, and solely 10% from algorithms.
search engine optimization groups are shopping for AI visibility trackers proper now, typically on a company card, which is exactly the sample Tony warns about. His Utility Take a look at has three gates, cleared earlier than any cash strikes. State the job in plain phrases with no vendor identify in it. Title the proprietor and the baseline. Set the one sign you’ll verify in 90 days, and determine upfront whether or not you’ll scale, repair, or cease.
“Observe our AI mentions” fails the primary gate. “Get us onto the AI shortlist when patrons ask for our narrowest class phrase” passes it, and Tony’s personal 1-in-43 is what a baseline appears to be like like.
How To Apply This To Your search engine optimization Technique
Tony closes with a five-move sequence. Right here’s how I’d adapt it for search engine optimization groups:
Add a suggestion charge to your scorecard, and report it first. Write 10 to fifteen prompts the way in which your patrons ask them, constructed across the class, the issue, and “who ought to I rent for.” Add two or three branded prompts. Run them throughout ChatGPT, Gemini, Perplexity, and AI Overviews, and monitor how typically you make the checklist and whether or not the outline is correct. Run it weekly, not as soon as, so that you’re studying a development reasonably than a snapshot.
Repair the file the fashions learn earlier than you chase the shortlist. Pull the sources the engines cite after they describe your model. Replace the pages you management, together with bios, LinkedIn, and the press web page, and request corrections from the aggregators and directories you don’t. After any rebrand or management change, assume the fashions haven’t heard.
Personal the slender phrase, then earn a spot on the lists. Choose probably the most particular method a purchaser describes your drawback and publish the canonical clarification below your model. Then discover the ten to twenty roundup and comparability pages the engines cite in your class and provides somebody a goal for getting on them. That’s a digital PR job, and it’s the job search engine optimization and PR ought to have been sharing all alongside.
Translate each AI visibility achieve into an owned relationship. Tie suggestion charge to signups, demo requests, subscribers, and pipeline, so your CFO sees the trustworthy finish of the spectrum reasonably than one other site visitors chart.
Identical Drawback, Totally different Galaxy
Tony and I haven’t labored on the identical firm or attended the identical assembly for 33 years. However, from a sure standpoint, we by no means stopped engaged on the identical drawback, which is getting the appropriate identify in entrance of the client earlier than the client goes wanting.
What’s modified is who builds the shortlist. My wager is that the groups nonetheless reporting rankings first will spend 2027 explaining a pipeline hole that their dashboards by no means confirmed them. The fashions are drafting your purchaser’s checklist proper now. Ensure you’re measuring whether or not you’re on it.
Extra Sources:
Featured Picture: patpitchaya/Shutterstock
