Quick Answer: SEO has always had two kinds of goals: hard goals like traffic and revenue that show up on a chart, and soft goals like visibility and perception that shape how leadership judges the program even when no click or conversion is attached. Traditional SEO gave you time to manage that gap, since a bad impression could usually be outrun by a good report before it hardened into an opinion. AEO removes that runway, because a single live AI answer now delivers the entire verdict in the moment it happens, with no report able to walk it back afterward. The fix is not a better reaction to that moment. It is not being surprised by it in the first place.
A CMO is on a call with a prospective agency, watching the pitch unfold with the usual mix of patience and skepticism. Partway through, she opens ChatGPT on the shared screen and asks it to name the top three tools in the category the agency has spent the last ten minutes describing. The company doesn't come up. Eli Schwartz recounts this exact moment in a recent newsletter, and the outcome was not a longer conversation about prompt variability or model grounding. The deal ended there. That SEO team no longer has a job, and the client cut the function entirely.
Nobody in that room cared that a different phrasing might have surfaced the brand, or that the same prompt run five minutes later could have returned a different list. The verdict had already landed, and it landed in about four seconds.
The Runway That Used to Exist
For most of SEO's history, perception moved slowly enough to manage after the fact. If a CEO had a bad feeling about search performance, based on a competitor's ranking, a tool that reported numbers incorrectly, or a term that mattered to them personally for reasons that had nothing to do with actual search volume, there was usually time to walk them back with a report, a trend line, or a well-timed win. The bad impression and the good data were racing on different clocks, and the data usually had enough of a head start to win.
That runway is what let SEO teams treat soft goals as a nuisance rather than a discipline. Chasing a ranking a board member cared about, or reporting Bing numbers nobody on the SEO team thought mattered, felt like a distraction from the real work, because the real work still had time to prove itself regardless of what any one person believed on any given day.
Why AEO Collapses That Runway
An AI answer does not wait for a quarterly report to form an opinion. A CEO who spent fifteen years half-trusting SEO dashboards can now ask an LLM about her own category and get an answer in the time it takes to read a text message, and whatever that answer says becomes her entire mental model of the program's performance, whether or not it holds up to scrutiny. There is no dashboard release cycle that runs faster than a curious executive with a phone.
This is what makes the scenario in Schwartz's newsletter land so hard. The SEO manager on that call reportedly responded with an explanation involving models, fan-out queries, and grounding, all technically reasonable points that had zero chance of mattering to anyone in the room. The soft goal had already failed in front of the people who controlled the budget, and no amount of technical accuracy afterward would undo four seconds of a blank answer on a shared screen.
Soft Goals Are Leading Indicators, Not Vanity Metrics
The instinct to dismiss soft goals as vanity metrics misses what they actually measure. A hard goal tells you whether the channel is working: traffic moved, revenue attached, a report you can defend in a board meeting. A soft goal tells you something a hard goal cannot: whether you matter to the specific people who decide your budget, your headcount, and your credibility, independent of whether any of that shows up in an attribution model.
Treating these as competing priorities is the mistake. They answer different questions that only look related from a distance, and a program that wins on one while ignoring the other is not half-succeeding. It is exposed in exactly the place that determines whether it survives its next budget review.
The Real Fix Isn't a Better Reaction
Most advice on this problem stops at damage control: know your audience, read the room, have a good answer ready when leadership asks an uncomfortable question. That advice is not wrong, but it treats the surprise itself as unavoidable, and in an AEO world the surprise is the entire problem.
The better fix is knowing the answer before anyone in the room asks the question. If a CEO is going to test your category on ChatGPT eventually, and at this point most of them will, there is no reason that test should ever catch your team off guard. The queries a leadership team is likely to run themselves are usually a short, predictable list: the company name, the two or three competitors that come up in every board conversation, and a handful of "best in category" phrasings that map to whatever language leadership actually uses, not necessarily the keyword list an SEO team built independently.
What This Looks Like in Practice
Build a leadership prompt list before you build anything else, separate from your standard keyword or category tracking. Sit down with whoever owns the relationship, ask directly what they would type into ChatGPT or Perplexity if they were checking on the category this weekend, and track that exact list on a standing schedule. Our earlier piece on B2B AI visibility covers in detail the mechanics of running this kind of audit.
Run that list across every platform leadership is actually likely to open rather than only the one your team tests by habit, since not every AI platform retrieves the same live information, and the version your CEO opens on her phone over the weekend may behave differently than the one your team uses for research on a Tuesday. Then report the hard goals and the soft goals in the same place, on the same cadence, so leadership never has to reconcile a traffic chart against a gut feeling on their own. When both numbers live in one system, a bad answer stops being a surprise that ends careers and starts being a gap you already knew about and were already closing.
Know the Answer Before the Question Gets Asked
The SEO manager in Schwartz's story did nothing technically wrong. The queries were probably imperfect, the model was probably inconsistent, and a rerun probably would have shown something different. None of that mattered, because the room had already decided, and no amount of being right after the fact was going to change a verdict that formed in four seconds.
The programs that will hold up under this kind of scrutiny are not the ones with the best explanation ready. They are the ones that already knew the answer before anyone thought to ask. If you want help building the kind of tracking that closes that gap before it becomes a Zoom call you can't walk back, the AEO Index tracks exactly this, and if you're ready to build the presence that makes the answer a good one, that is the work our AEO and GEO team does every day.