THE RECOMMENDATION LAYER · MAY 27, 2026 · 9 MIN READ

The Silent Vetting: The Shortlists You Never See Yourself Lose

AI assistants build your buyers' shortlists before any contact happens. Being left off produces no signal at all. Who you lose first, and how to find out.

By Jax Baker
TL;DR — WHAT TO REMEMBER

Originally published on Jax5d in January 2026 as part 4 of Jacquie Baker’s five-part series on the first 1,000 days of mainstream AI. Adapted and expanded for Probably Genius.

The silent vetting is the shortlist an AI assistant builds for your buyer before you know the buyer exists, and losing your place on it produces no signal at all: no rejection email, no lost bid, no dip in your analytics. The buyer asks their assistant who to hire, the assistant reads the record on your market, and it hands back one name, maybe two. If yours is not one of them, the loss never registers anywhere you can see.

Losing in business used to be loud. You pitched, and they said no. You bid, and someone else won. You had a metric for failure: rejection.

The new failure has no metric. It has no notification. It is simply silence.

That silence is easy to dismiss as a futurist’s worry, so put a number next to it. Adobe Analytics measured traffic arriving at U.S. retail sites from generative AI sources growing 1,200 percent between July 2024 and February 2025, and in Adobe’s survey of 5,000 U.S. consumers, 87 percent said they were more likely to use AI for larger or more complex purchases.

Larger and more complex is precisely what a professional service is. The higher the stakes, the earlier the machine enters the conversation.

Our co-founder Jacquie Baker published the original version of this argument in January 2026, from an unusual seat: she spends her days wiring up AI agents for clients, workflows that scrape, synthesize, and decide what is relevant. This is not a commentator’s take on what AI might do to your pipeline. It is a builder’s report on what this class of machinery already does, written from inside it.

The Death of the Menu

For twenty years the internet ran on what the original essay called the Menu Model. You searched for “best brand strategist.” Google handed you a menu of ten blue links. You opened five tabs, read the About pages, squinted at the photos, and built your own shortlist. You, the human, did the processing.

That era is ending in front of us. Now a founder asks her assistant a question no keyword tool has ever seen: “Who is the best person to help me fix my brand positioning before a Series B raise? I need someone who understands the space but isn’t a bro about it.”

The assistant does not hand her a menu. It does the reading for her. It scans the tabs in milliseconds, checks the About pages, cross-references the profiles, and hands back one name. Maybe two.

If you weren’t the answer, you didn’t just lose the deal. You were never even in the room. You weren’t rejected. You were omitted.

The selection behind that answer is stricter than most owners assume. When Ahrefs analyzed 1.4 million ChatGPT prompts, it found the model cited only 49.98 percent of the URLs it had already retrieved. Half of everything that even reaches the machine’s desk never makes the answer.

The vetting is real, it is mechanical, and it runs every day in your category, in conversations you never see, with or without you.

One honest caveat, because the honest version is the one worth publishing: no one measures these shortlists from the outside perfectly, including us. Any single answer from any single engine is a data point, not a verdict. But the direction of the behavior is not in dispute, and waiting for perfect measurement is how the window closes.

The Failure With No Error Message

Every failure you have ever managed had an event attached to it. The proposal declined. The call that never got returned. The competitor’s press release. Painful, but countable, and what gets counted gets fixed.

The silent vetting produces a new kind of loss: one with no event. The buyer who never called did not bounce off your website; they never reached it. The bake-off happened inside a chat window, in the time it took the machine to compose a paragraph.

Your analytics cannot show you a visit that was never referred, and your CRM cannot log an inquiry that was never made. This is why AI visibility is a different scoreboard from traffic: the losses happen upstream of every number you currently watch.

The original essay carried a moment we have seen repeat on audit calls, retold here with the owner’s details kept private. Partway through a review of who the engines actually cite in his space, the owner went quiet. Then he started searching his competitors on screen, companies he had never heard of, clicking through their sites and their client lists. “I tendered on that job,” he said. “And that one.”

Work he had bid on, work he was qualified for, going to firms that showed up where he didn’t. Not because they were better. Because they were findable. Mapping exactly that, who currently occupies the answers a market’s buyers hear, is why we build a Growth Map before we build anything else.

Sit with how strange his position was. He was not failing. Referrals still came. The book of business looked healthy. The losses were real money all the same; they just belonged to a category his instruments could not display.

To a human who meets you, you’re obviously credible. To AI, you’re a blank slate. Nothing in his world was built to tell him those two sentences could both be true.

The Buyers You Lose First

Here is the part that should genuinely bother you: the leads lost to silent vetting are not average leads. They are the best ones.

Watch how an AI-first buyer actually moves. Someone asks their assistant about the wet patch near the toilet they cannot trace. The assistant walks through likely causes, explains when this is a DIY job and when it needs a professional, gives a realistic cost range. Then, finally: “Can you recommend someone in my area?”

By the time that person contacts a plumber, they know what is probably wrong, they know they need help, and their expectations on price are already set. They’re not shopping. They’re confirming.

The original essay named this person the Researched Buyer, and the behavior is now measurable at scale. In Adobe’s data, visitors arriving from generative AI sources viewed 12 percent more pages per visit and bounced 23 percent less than visitors from other channels; they were still slightly less likely to convert, but that gap collapsed from 43 percent to 9 percent in seven months, which is the sound of a behavior hardening into a habit. These buyers are not coming back to the menu.

When the machine has already diagnosed the problem, set the expectations, and matched the person to a provider, most of the sales work is done before any website gets visited. Lose the silent vetting and someone else receives that pre-sold buyer. Someone who might have half your experience but twice your presence.

Which raises the only question that matters: why would an excellent firm lose to a thinner one? The answer from inside the systems is blunt. Most experts aren’t invisible because they lack expertise. They’re invisible because their expertise is trapped.

The keynote from last year. The podcast where you explained your whole methodology. The case-study PDF buried three clicks deep behind a form. The client calls where you say the brilliant things you never write down. The machine cannot attend your talks or sit in on your calls. It can only see what’s structured, published, labeled and connected.

We named this pattern the Black Box Effect, and the original essay put its consequence in one line: if your expertise lives in the black box, AI has nothing to recommend.

How to Find Out, and What Changes It

You do not have to take any of this on faith, and you should not. The vetting is observable. Write the three questions your buyers actually ask, put them to three AI assistants word for word, and read the nine answers that come back: who got named, who got cited, and whether you appear anywhere in the room where your market’s decisions now get made.

We published the full method as the ten-minute test, free, because the fastest cure for disbelief is nine answers about your own market with your name missing from all of them.

If the test stings, the way back is not louder marketing, and it is not gaming anything. The machines are running an evidence check, and the check has a principle behind it that carries this entire series: AI can’t cite what you won’t claim.

The work is getting what is already true about you onto the record in forms a cautious machine can verify: positions published under your name, your methodology written down and titled on a page you own, one coherent identity across every surface the engines consult, and the proof counted rather than gestured at. Coherence is what the machines read. Coherence creates visibility, and visibility is what turns the silent vetting from a leak into a pipeline.

In our own client work, that mapping is the whole first act: structured interviews, case studies, named frameworks, specific proof points, compiled until the record finally matches the practitioner. Not keyword stuffing. Getting the brilliance out of your talks, your calls, and your head, structured so the machines can read it and connected so they can trust it.

The expert is the genius in that equation; we are the translators. If a buyer asks their assistant for someone kind but firm who knows your market, the evidence of your kindness, your firmness, and your market needs to exist somewhere a machine can find it.

And the quiet, strange as it sounds, is the good news. Right now, in category after category we test, the silent vetting has no confident winner: the answers are full of directories and roundups, not rivals. Most of your competitors are still fighting for scraps on the search results page while the real decision moved rooms.

Very quiet markets are unclaimed markets. The question is whether you use the window to get on the record, or keep waiting in the silence.

Want to Learn More?

Probably Genius was built by Jacquie Baker and Christopher Shaw, who spent more than 40 combined years translating what makes an expert the best in the room for the audiences that decide who gets chosen. Jacquie has been publishing the thinking behind this work on her Substack, Jax5d, since 2025; this essay adapts the fourth part of her series on what the first 1,000 days of mainstream AI changed for experts and the businesses built on their judgment.

Frequently asked questions

What is the silent vetting?
The silent vetting is the shortlisting an AI assistant performs for a buyer before any human contact happens. The buyer describes their problem in conversation, the assistant reads the available record on the businesses in that market, and it recommends one or two names. Businesses left off that shortlist receive no signal of the loss: no inquiry, no rejection, no analytics event. The term comes from Jacquie Baker’s January 2026 essay of the same name, part 4 of her series on the first 1,000 days of mainstream AI.
Why don’t these losses show up in my analytics?
Because they happen upstream of everything your analytics can observe. A buyer who is vetted out never clicks through to your site, never fills a form, and never calls, so there is no visit to attribute and no lead to mark lost. Traffic and ranking dashboards measure the behavior of people who reached you; the silent vetting decides who reaches you. That is why testing AI answers directly, rather than reading traffic reports, is the only way to see it.
Who is the Researched Buyer?
The Researched Buyer is a prospect who has already worked through their problem with an AI assistant before contacting anyone: likely causes, realistic costs, when professional help is warranted, and who to call. They arrive pre-qualified, decide quickly, and negotiate less, because the qualifying conversation already happened. Adobe’s 2025 data shows visitors from generative AI sources browsing 12 percent more pages with 23 percent lower bounce rates than other traffic, consistent with buyers who arrive already convinced they are in the right place.
How do I find out whether AI is leaving my business off buyer shortlists?
Run the ten-minute test: write the three questions your buyers actually ask, put them word for word to ChatGPT, Gemini, and Perplexity in fresh chats, and score the nine answers for whether you were named, described accurately, and recommended. If you want the why behind the result scored properly, a full diagnostic runs the same recommend-test as a panel across five engines inside a human-verified examination of eight zones, and hands you the fix path in priority order.

CITATIONS

  1. “The Silent Vetting: Part 4” (Jax5d, January 2026). The original essay this article adapts and expands: the Menu Model, the omission frame, the Researched Buyer, and the audit-call story retold above, published by Probably Genius co-founder Jacquie Baker. jax5d.substack.com
  2. “Adobe Analytics: Traffic to U.S. Retail Websites From Generative AI Sources Jumps 1,200 Percent” (Adobe, March 2025). The measurement behind this article’s adoption figures: 1,200 percent growth in AI-sourced traffic from July 2024 to February 2025, 12 percent more pages per visit, 23 percent lower bounce, the 43-to-9 percent conversion-gap collapse, and the 5,000-person survey (87 percent more likely to use AI for larger or complex purchases; 92 percent said it improved the experience). blog.adobe.com
  3. “Why ChatGPT Cites One Page Over Another (Study of 1.4M Prompts)” (Ahrefs, April 2026). Large-scale study of ChatGPT citation behavior: only 49.98 percent of retrieved URLs were cited, with title-to-question relevance the strongest survival factor. ahrefs.com
WRITTEN BYJacquie ("Jax") Baker

Founder of Probably Genius, an AI visibility firm helping professional service brands become the named answer in AI search. Nearly two decades across technology, digital strategy and branding — including 1,000+ digital projects through her previous agency — now focused on making experts visible, verifiable and recommendable to AI. Let's talk →

Related reading

Read the method.
Then see it run on you.

The 109-point diagnostic is this library, applied to your business. Free, about an hour to present, no obligation.

See where you stand →

A real diagnostic, not a sales call in disguise.