Attention got cheap. Trust didn't.

Every era of marketing is a bet on whatever is scarce. For the past decade the bet was attention: content was expensive to make, so the brands that could afford more of it, distributed louder, won. AI voided that bet in roughly two years. Content is now abundant and close to free, and nobody competes for an abundant thing. In May 2026, from the stage of its marketing symposium in London, Gartner gave the new scarcity a name: trust scarcity. The claim, stripped to plain words: the contest is no longer whether people see you. It is whether anyone, human or machine, believes you.

This is an analysis, and it stands on three studies that measure three different things. Gartner measures how much consumers trust big brands. BCG measures how much consumers trust AI's answers. Edelman measures how much people trust institutions. They are not interchangeable, and we will not blend them. Read side by side, they describe one movement: trust is draining out of the loud channels and pooling in the machine's synthesis.

Section OneWhy attention was ever the prize

Attention was never valuable in itself. It was valuable because it was rationed. Producing a decent article, a video, a campaign took money and skilled hours, so the supply of content stayed roughly proportional to marketing budgets. And human attention, fixed at the same waking hours it has always been, became the bottleneck everyone priced. The entire architecture of digital marketing, from the ad auction to the SEO industry to the social feed, is machinery for allocating that bottleneck.

Generative AI broke the proportionality. The cost of producing plausible content has collapsed toward zero, and the volume is following the price. Gartner's analysis draws the consequence cleanly: when content becomes abundant and low-cost, what is scarce changes, and with it, what brand building means. In Gartner's framing, brand building shifts from buying exposure to the deliberate design of trust signals: the verifiable marks that let a buyer, or a machine reading on the buyer's behalf, decide you are real.

Gartner, A. Abatangle, "In the AI Era, Trust Scarcity Is Rewriting the Rules of Brand Growth," May 12, 2026.

Content got cheap. Belief didn't. The competition moved to where the price stayed high.

Section TwoA ten-point slide, and a sharper sentence

Gartner brings a number. In its survey of 1,000+ UK consumers, fielded August–September 2025, 60 percent said they trust big brands, down from 70 percent in 2021. One vendor's survey, one country, one wave: we flag that plainly, because the number will travel further than its caveats. But the direction matches what every other instrument in this piece is picking up, and a ten-point slide in four years is not noise a brand strategy can ignore.

Gartner, A. Abatangle, "In the AI Era, Trust Scarcity Is Rewriting the Rules of Brand Growth," May 12, 2026 (survey of 1,000+ UK consumers, Aug–Sept 2025).

The sharper finding is not the slide. It is what Gartner says trust now does. Trust, in their analysis, no longer influences just preference. It influences discoverability. That is a category change. Preference decides which of the options in front of a buyer wins. Discoverability decides which options are in front of the buyer at all. When AI assistants assemble the shortlist, the brands the machine's sources corroborate get shown; the rest are not rejected, they are simply never surfaced.

Section ThreeThe counterweight: what consumers trust instead

Now the other instrument, measuring a different thing entirely. BCG's Global Consumer Radar (9,000+ respondents across nine countries, surveyed in waves in September 2023 and October 2025) finds that more than 60 percent of consumers trust AI-generated results to guide their purchase decisions. And they are acting on it: BCG measured the use of generative AI for shopping growing 35 percent between February and November 2025.

BCG, "Consumers Trust AI to Buy Better. Brands Must Adapt.," 2026 (Global Consumer Radar, Sept 2023 and Oct 2025 waves).

Hold the two findings apart, because they measure different objects: Gartner asked UK consumers about brands; BCG asked a nine-country panel about AI's outputs. But set side by side, they form the strategic picture of the decade. Trust in the sellers is falling. Trust in the machine that summarizes the sellers is high and rising. The buyer has not stopped trusting. The buyer has relocated their trust one layer up, from the source to the synthesis.

Consumers trust brands less and the machine's summary of brands more. The machine is becoming the trust intermediary.

There is a third instrument worth placing on the table, strictly as context, because it measures institutions rather than brands. The 2026 Edelman Trust Barometer finds 37 percent of respondents naming the growing use of generative AI among the top forces that have reshaped trust in institutions over the past five years. It also finds 54 percent of low-income respondents (44 percent of middle-income) believing generative AI will leave people like them behind. People lean on the machine's answers and are uneasy about the machine's rise, at the same time. Both are true, and neither cancels the other.

Edelman, "2026 Edelman Trust Barometer," 2026 (institutional trust, not brand trust; context only).

Section FourBeing trusted by the machine

If the machine is the new intermediary of trust, then the new brand-building question is uncomfortable and precise: what does it take to be trusted by the machine? Not liked by an algorithm. Trusted by a system that reads the public record about you and decides what to repeat.

The answer looks nothing like an ad budget. An AI assistant assembling a recommendation weighs what it can verify: whether the facts about you agree with each other across every surface where you appear; whether parties other than you (reviewers, press, directories, forums, industry lists) say the same thing you say about yourself; whether your claims come with checkable specifics or only adjectives. Self-description is the one signal the machine discounts most heavily, for the same reason a judge discounts a defendant's own character reference.

Trust is now earned in sources you don't own.

This is what Gartner's phrase "intentional trust-signal design" cashes out to in practice. The decade of attention rewarded volume on your own channels. The era of trust scarcity rewards corroboration on everyone else's: the review you cannot edit, the third-party mention you did not write, the consistent, structured, verifiable record that lets a machine conclude, cheaply and repeatedly, that you are who you claim to be.

Section FiveWhat to do with this

The practical program follows from the mechanism, and none of it is exotic. Make your facts agree. Name, offer, prices, claims, dates: identical on your site, your profiles, the directories, the registries. Contradiction is the cheapest distrust signal a machine can detect, and it detects all of it.

Invest where you have no edit rights. Reviews, independent mentions, citations, customer forums. One corroborating voice you don't control outweighs pages of self-description you do.

Attach receipts to claims. "Trusted by thousands" is noise. A number, a name, a date, a checkable reference is signal. Write for a reader who will verify, because the machine reading you does.

Stop budgeting for volume. Ten more AI-written posts on your own blog add to the abundance the machine filters out. One earned, specific, third-party-visible proof point adds to the scarcity it filters for.

In ClosingThe same thesis, one layer deeper

Everything we publish argues one claim: machines can only act on what they can clearly read, outside your company and now inside it. Trust scarcity is that claim wearing economic clothes. When content became free, the machine stopped rewarding whoever wrote the most and started rewarding whoever could be verified: consistent, corroborated, legible in the sources it consults before it ever mentions your name. The decade of attention asked: can they see you? The era of trust asks a harder question: when the machine checks, does your story hold?

You can measure how visible, and how corroborated, your brand currently is to the machines at /signal-index/, or write to us at /contact/.