Afternoon BriefAI Search & Discovery

Consumer Trust in AI Search Dropped 28 Points While Usage Surged 70%. The Brands That Win the Validation Cycle Are Already Decided.

Fractl's Q2 2026 study shows AI search trust fell from 82% to 54% while adoption climbed 70%. The trust gap creates a validation cycle where earned authority determines which brands survive AI discovery.

Jaxon Parrott
Jaxon ParrottJun 18, 2026

Fractl's Q2 2026 study surveyed 1,008 consumers and 150 marketers and found a paradox that should restructure how every founder thinks about brand discovery: consumer trust in AI search dropped from 82% to 54% in one year while usage surged 70%. The gap between those two numbers is where the next wave of brand advantage gets built or lost.

I have spent the last eight years placing brands in front of the audiences that decide whether those brands matter. I say that because the data Fractl just published confirms something I have been watching happen in real time across our client portfolio: the brands with third-party earned authority are winning AI discovery. The brands without it are invisible at the exact moment the buyer is checking.

The Trust-Usage Paradox Is Real, and the Numbers Are Stark

Here is what Fractl found. In 2025, 82% of consumers said AI search was more helpful than traditional search. One year later, that number fell to 54%. The share of consumers who actively rate AI as less helpful than regular search grew 6x, from 3% to 17%.

At the same time, 70% of consumers said they are using AI search tools more than they did last year. Only 3% said they use them less.

This is not a contradiction. It is a filter. Consumers adopted AI search for convenience. Then they started catching the errors. 27% of marketers in the same study said their brand had been misrepresented in AI answers. 14% reported real business impact from those inaccuracies, including lost sales and PR damage.

The result: people use AI search more, but they trust it less. That trust deficit changes what happens after the AI answer appears.

What Happens When Trust Drops but Usage Stays

When consumers trust the tool less, they validate more. Fractl's data shows buyers now check an average of 2.4 platforms before making a purchase decision. That number holds across Gen Z, Millennials, and Boomers.

Here is the validation hierarchy. When asked which platform they trust most for product recommendations:

  • Google search results: 39%
  • Reddit: 15%
  • AI tools (ChatGPT, Perplexity, Gemini): 14%
  • Review sites: 11%
  • Friends and family: 11%

Google still leads AI tools three to one for purchase trust. Reddit, a platform built entirely on earned human opinion, outranks every AI tool in the category.

This is the validation cycle. The AI answer is the starting line. The buyer then runs a second search, checks a review site, or asks a forum. If your brand shows up in the AI answer and has earned corroboration in those validation layers, you win the cycle. If you only show up in the AI answer with no third-party backing, the buyer moves on.

The Generation Gap Is Not What You Think

The conventional read is that younger consumers trust AI more. The data says the opposite.

Baby Boomers are more likely to find AI search helpful (63%) than Gen Z (47%). And when asked whether they would reduce trust in a brand that uses AI heavily, 54% of Gen Z said yes. Only 33% of Gen X and 32% of Boomers said the same.

WordPress VIP's June 2026 survey found that 60% of U.S. consumers say "AI" in brand messaging is a turnoff.

The generation most native to AI tools is also the most suspicious of brands that rely on them. This is not an anti-technology sentiment. It is a quality filter. Gen Z can spot generated content because they generate content themselves. They know what the machine sounds like, and they penalize brands that sound like it.

The Zero-Click Reality Makes Earned Authority Non-Optional

Similarweb's 2026 data shows 68% of Google searches now end without a click. On desktop, it is 80%. When an AI Overview appears, click probability drops from 15% to 8%.

Google AI Mode has passed 1 billion monthly users. But its referral rate sits at 1.6% to 2.5% of queries versus traditional Google at 17% to 19%.

The clicks that do survive are worth more. ChatGPT referral traffic shows an average session time of 15 minutes versus 8 minutes from Google, 12 pages per session versus 9, and a 7% conversion rate on transactional sites versus 5% from Google search.

So the math is: fewer total clicks, but the clicks that survive the AI filter convert better because the buyer already read the summary and chose to go deeper. The question is whether your brand is the one they choose to go deeper on.

That choice happens in the validation cycle. The buyer saw you in the AI answer, checked whether you have earned authority backing the claim, and then either clicked through or kept scrolling.

What the Brands Getting This Right Are Actually Doing

50% of marketers in the Fractl study reported decreased organic traffic since AI Overviews launched. Only 12% say they have a GEO strategy with measurable results. The gap between knowing something changed and doing something about it is enormous.

A Princeton GEO study found two tactics that move the needle on LLM citation rates: direct expert quotations improve citation rates by 41%, and quantified statistics improve them by 32%.

Those are not SEO tactics. Those are earned authority tactics. An expert quote only exists if you have placed a human expert in a credible publication. A quantified statistic only lands if you have published original research or been cited in someone else's.

This is Machine Relations at its core: earning the citations, mentions, and third-party references that AI engines use to decide which brands are worth surfacing. Not optimizing a page. Building the source architecture that makes the machine trust you before the consumer ever asks the question.

The 49% of marketers who monitor LLM visibility in 2026, up from 22% in 2025, are at least watching. But watching is not the same as building. The brands that will own the next twelve months are the ones converting earned media placements into extractable proof: named expert quotes, specific data points, third-party corroboration that the AI engine can verify across multiple sources.

FAQ

How much has consumer trust in AI search declined?

Fractl's Q2 2026 study of 1,008 U.S. consumers found trust dropped 28 percentage points in one year, from 82% finding AI more helpful than traditional search in 2025 to 54% in 2026. The share actively rating AI less helpful grew 6x.

Why do consumers use AI search more if they trust it less?

Convenience drives adoption while experience erodes trust. 70% of consumers increased AI search usage, but 40% now say they would trust a brand less if it used AI heavily, up from 20% in 2025. Consumers use the tool for speed, then validate elsewhere.

The validation cycle is the multi-platform check buyers run after seeing an AI answer. Consumers check an average of 2.4 platforms before purchase decisions. If a brand appears in the AI answer but has no earned third-party corroboration on review sites, search results, or forums, the buyer skips it.

Additional source context

Continue Exploring

The most relevant next read from elsewhere on AuthorityTech.

BlogJun 18
Earned vs. Owned AI Citation Rates: What the 2026 Data Actually ShowsExecution playbook
BlogMar 20
What Is Machine Relations? The Discipline Behind GEO, AEO, and AI Search VisibilityExecution playbook
Glossary
AI Brand AuthorityRelated glossary concept
ResultsJun 3
How GEO Is Replacing Traditional SEO for Ecommerce — And Which Platforms Are Already ThereLive proof