A B2B buyer used to open ten to twenty browser tabs. They compared vendors on G2 and Capterra and built a spreadsheet by hand. Now that same buyer opens Claude or ChatGPT instead. They describe their needs in a paragraph and get a comparison table back in seconds. Research that once took days now takes minutes. Without an answer engine optimization strategy, a vendor may never make the model’s list at all.
That shift drove the season four opener of Misadventures in Marketing, AMA SF’s podcast by Steve Haney and Peter Farago. Together they explored what happens when an algorithm, not a procurement officer, writes the first draft of the shortlist.
Why Is AI Replacing the Traditional Buyer Journey?
Farago framed the shift around a familiar B2B idea: low-consideration versus high-consideration purchases. Complex products have always demanded more research. What has changed is where that research happens. Farago cited research putting the average Google query at about five keywords. A prompt to an AI assistant, by his numbers, runs closer to twenty or twenty-five words. Buyers give the model far more context, and they expect a tailored answer back.
That longer, conversational query changes what marketers need to produce. Teams can no longer just anticipate search terms. They have to anticipate the actual question a buyer will type. Farago calls this the “invisible RFP.” No formal document changes hands, yet the AI still scores every vendor against real buying criteria. It also casts a far wider net than any person could. “AI can look at a hundred products and compare them,” Farago said.
What Does It Take to Survive the AI Shortlist?
Farago compared the first round to Squid Game. Far more contestants enter, since the AI can pull in a hundred vendors, and the cuts come fast. He named five tests a company must pass to stay in the game.
– Eligibility: does the model recognize your category and the terms buyers actually use?
– Fit: does your content answer every dimension a buyer is comparing?
– Extractability: can the model locate that information and map it to its table?
– Credibility: do third parties, not just your own site, back your claims?
– Consistency: do your website, docs, and reviews tell the same story?
“Does the machine even understand the category we’re in?” Farago asked. His own company calls itself an AI SRE, short for site reliability engineering. The team debated whether “DevOps agent” was the better label before settling on its terms. That choice matters in any category, because a model matches vendors to the buyer’s vocabulary. Pick the wrong words, and your company may simply not appear.
Where Should Answer Engine Optimization Efforts Go First?
Haney raised a fair worry about gaming the system. He recalled the offshore link farms that once tricked Google’s algorithm into ranking dubious sites. Will marketers now load their sites with tricks to fool the bots onto a shortlist? That question stayed open, but many of Farago’s tips favor sources that are hard to fake.
Wikipedia tops that list. Farago called it “very still pristine and crowdsourced and heavily edited and monitored.” Someone other than the company has to write the page, and every claim needs a credible citation. Reddit conversations and review sites such as G2, TrustRadius, and Capterra also carry weight for many categories.
On your own site, Farago recommended an expansive FAQ section. He also urged marketers to stop shying away from competitors. Publish a big, detailed, accurate table comparing your product against named rivals. A roundup-style blog post can do the same job. Since the agent is hunting for exactly that kind of structured contrast, give it one you wrote.
Is There One AEO Playbook for Every Category?
Farago was blunt on this point: no single formula exists. Every model weighs sources differently, and the mix shifts by product category. Wikipedia might dominate for one category while Reddit or your own website drives another. So marketers have to test their own category and right-size the effort to fit. “There’s not a one size fits all solution,” he said.
Does a Human Still Matter in an AI-Led Sale?
Even after a company survives the algorithmic cut, people still close the deal. Haney argued that the bot will soon handle **60 to 70 percent** of the technical analysis. Buyers will arrive at the sales call after running scenarios across several vendors. That flips an information edge that used to belong to the account executive.
Yet no AI will take the blame for a failed purchase. “A human’s gotta be there to be accountable for the purchase,” Haney said. Your champion inside the account wants a promotion, a raise, or better numbers. Make that person look good, or they stop championing you. Farago added a practical caveat: you only reach that human stage after making the shortlist.
Haney invoked Intuit founder Scott Cook’s definition of brand: “Your brand is what people say about your company when you’re not in the room.” That idea grows sharper once an AI agent joins the room too. Haney described the agent as a new member of the buying committee, one that is “super intelligent” and “very quick.” Farago liked that framing and used it to explain how marketers should measure success.
Haney compared the two audiences to Star Trek’s Spock and Kirk. Spock wants clean data and verifiable claims, ideally in markdown files the bots can parse. Kirk wants a relationship: eye contact, drinks, a great conversation at a party. Haney’s verdict was simple. “If you’re consistent between both, that’s really, I think, gonna be the secret sauce to success in this space.”
How Should Marketers Start Measuring AI Visibility?
Farago offered a concrete first step: shift from keyword lists to buyer prompts. Build a dozen realistic prompts a target buyer might type. Organize them around recurring problems, requirements, and buying criteria. Then ask real buyers what they typed, and run the research process yourself.
Repetition matters, because one query tells a marketer very little. Run the same prompt many times with small variations, then count what comes back. Which companies appear, and does yours make the cut? How does the model describe you, and which competitors does it compare you against? Which sources does it cite? Those citations reveal where the model grounds its answers, and therefore where you need to show up.
Tracking platforms already automate this work, and Farago named Profound as one example. Still, any team can start by hand, one prompt at a time. Farago treats it as always-on hygiene rather than a one-time project, since buyers keep shifting their research into these tools.
The desire to buy hasn’t changed, Farago stressed, and neither has the marketer’s job. What changed is the path between a first spark of interest and a signed contract. More of that path now runs through a model that can weigh a hundred vendors without getting tired. Winning the humans still closes the deal, but only for the vendors the model let through.
Listen to the full podcast on Spotify and Apple Podcasts.
Check out summaries from other episodes:
- Career Positioning in the Strangest Job Market We’ve Ever Seen
- The Marketing Playbook Is Dead. The Job Is the Playbook Now
- GEO, Dark Funnel, Prompt Slop: An AI Marketing Fluency Check
Misadventures in Marketing is a weekly podcast by the AMA San Francisco chapter. Veteran Silicon Valley marketing execs Peter Farago and Steve Haney explore the messy, rewarding, and occasionally absurd world of high-tech marketing – especially in early-stage startups. Each episode covers real-world challenges, trends, and lessons from the front lines.
Comments are closed.