What new AI search data reveals about visibility and trust

Trust successful AI hunt is declining, consumers are validating accusation crossed much platforms, and AI visibility is progressively tied to marque authority alternatively than accepted SEO metrics.

Those are among the cardinal findings from caller investigation by Fractl and MCP, presented by Fractl cofounder Kelsey Libert astatine SMX Advanced. The study offers a elaborate look astatine really consumers measure AI-generated answers, which signals power AI recommendations, and wherever brands are falling short connected governance and disclosure.

After the conference, I said pinch Libert to excavation deeper into the data. Our speech covered the increasing spot spread successful AI search, the domiciled of earned media and entity authority successful AI visibility, and why galore organizations are still unprepared for the operational challenges AI introduces.

The honeymoon is over

The header uncovering from Libert’s investigation is difficult to ignore. In 2025, 82% of consumers recovered AI hunt much adjuvant than accepted hunt results. By 2026, that fig had dropped to 54%, a diminution of 28 percent points successful a azygous year. The skeptic campy grew sixfold complete the aforesaid period.

I asked Libert what she thinks is driving that erosion.

• “Hallucinations. Initially, AI was a frictionless instant-answer instrumentality that felt superior to Google’s crowded SERPs. As group mislaid spot successful AI answers and had to put successful much effort to validate them, that instant gratification disappeared, and the helpfulness people dropped dramatically,” Libert said.

She isn’t wholly pessimistic astir wherever things are headed, though.

• “AI is connected an exponential betterment scale, truthful I expect this number to restabilize complete the adjacent twelvemonth arsenic group study really to refine their prompts and engineers amended the technology,” Libert said.

That restabilization whitethorn travel sooner than expected. Libert pointed to a June 5 CNN study covering Anthropic’s warnings that AI whitethorn soon beryllium tin of improving itself without quality intervention. Whether that accelerates the betterment of user spot aliases deepens interest astir AI reliability remains to beryllium seen.

In the meantime, consumers are hedging. The investigation recovered that buyers now cheque an mean of 2.4 platforms earlier validating a purchase, a shape accordant crossed Gen Z, millennials, and boomers.

Google still leads AI devices 3 to 1 for trusted merchandise recommendations, commanding 39% of user spot versus 14% for AI tools. Reddit, astatine 15%, sits betwixt them.

Organic visibility is diversifying

For SEOs worried astir the erosion of integrated traffic, Libert’s framing offers a much nuanced image than the emblematic punishment narrative. About 50% of marketers study postulation declines since AI Overviews launched, and 61% straight blasted AI tools.

At the aforesaid time, 57% spot postulation maturation from societal platforms, including TikTok, Reddit, and YouTube, and 40% spot maturation from AI assistants specified arsenic ChatGPT and Perplexity.

The transmission representation Libert presented astatine SMX is worthy understanding. 

• Google remains ascendant astatine 84.8 cardinal visits, serving chiefly arsenic an intent-capture engine. 
• YouTube and Instagram/TikTok together grip marque discovery. 
• ChatGPT and Gemini are utilized chiefly for investigation and learning. 
• Facebook and Reddit service human-validation functions. 

Search isn’t disappearing. It’s fragmenting, and brands that optimize for only 1 of these channels are leaving important visibility connected the table.

The GEO hierarchy: Table stakes, precocious risk, and the moat

Libert’s investigation categorized generative motor optimization strategies into 3 tiers, and the distinctions matter for really marketers should allocate effort.

The astir commonly utilized maneuver is FAQ optimization, employed by 49% of marketers. Libert calls this precocious risk, and the logic is straightforward. 

• “The high-risk class is based connected really easy AI tin replicate that content, and wide manufacture FAQs are typically beautiful easy for AI and your competitors to produce.”

FAQ strategies tin beryllium strengthened by layering successful proprietary data, taxable matter expertise, and unsocial insights, but connected their own, they connection small defensibility.

Table-stakes strategies see building marque mentions (43%), establishing topical authority (36%), and implementing system information (30%). These are basal but not sufficient.

The moat, arsenic Libert describes it, consists of original information and proprietary studies (35%) and integer PR (24%). These are harder for AI to replicate, which is precisely what makes them valuable. 

• “LLMs and Google request caller contented to propulsion into timely and communal RAG queries. Beyond that, information publicity and integer PR efforts thief summation your brand’s entity authority by helping you gain marque mentions crossed influential sources crossed the web,” Libert explained.

The awesome spot down this attack is backed by Ahrefs investigation analyzing astir 75,000 brands crossed ChatGPT, AI Mode, and AI Overviews. 

Branded web mentions and YouTube impressions showed the strongest correlations pinch AI visibility, ranging from 0.50 to 0.74 connected the Spearman scale. 

Backlink count and advertisement walk fell successful the weakest tier, beneath 0.30. The applicable implication, arsenic Libert put it, is that “AI systems reward marque beingness and mentions much than accepted SEO standard metrics.”

I asked Libert what she would do first if building marque mentions from scratch. Her reply was tactical and specific:

• Use Semrush aliases Ahrefs to place high-authority, niche-relevant publishers that person covered competitors, build relationships pinch those journalists, and transportation concepts that capable gaps successful their sum utilizing proprietary information aliases taxable matter expertise.
• Use SparkToro to place the publishers, YouTube channels, and podcasts a target assemblage really consumes, past prioritize earned placements successful those venues.
• Use Reddit to study what performs good successful applicable subreddits and lend substantive commentary aliases content, not promotional noise.

• “Earned media ever comes backmost to 1 principle. Focus connected creating fresh, educational, actionable, valuable, and newsworthy manufacture insights, and repurpose that contented crossed the channels of power for your target market,” LIbert said.

---

The spot spread brands are ignoring

One of the astir striking information points successful Libert’s position involves the spread betwixt what consumers expect and what brands really do. 

Between 84% and 91% of consumers opportunity they want AI labeling crossed each contented formats, including text, video, audio, and images. Only 20% of organizations opportunity they ever disclose their AI use. One-third ne'er disclose astatine all.

Libert doesn’t deliberation consumers are simply opposed to AI successful marketing. The interest is much specific. 

• “Consumers aren’t wary astir AI usage successful marketing. They’re concerned erstwhile brands usage it for their full trading workflow pinch zero checks and balances.” 

She cited the investigation uncovering that astir half of marketers admit to not fact-checking their AI-produced contented and that 48% opportunity AI makes their activity faster but much average.

Dove came up arsenic 1 of the fewer brands actively positioning against what Libert calls “AI-dominant trading slop,” though she acknowledges the section is still mostly treating disclosure arsenic a compliance checkbox alternatively than a marque signal.

There is besides a deeper mobility astir wherever AI assistance ends and quality productivity begins, 1 the manufacture hasn’t settled. 

• “There’s immoderate statement connected erstwhile it needs to beryllium disclosed, since plentifulness of group usage it much arsenic an adjunct successful the imaginative process vs. having it do everything end-to-end,” Libert said. “In that case, who genuinely owns the creativity, and what level of disclosure is required? That’s still up for debate.”

The inverted pyramid problem

The descent that drew the astir guidance successful the SMX room showed really marketers are allocating their quality reappraisal clip for AI-generated content. Editorial reappraisal claimed 72% of attention. Voice and reside reappraisal claimed 62%. Fact-checking fell to 54%, plagiarism and ineligible reappraisal to 42% each, and bias information to conscionable 27%.

Libert called this an “all of the above” problem, spanning training gaps, workflow fragmentation, and activity priorities. 

• “When labor can’t support up pinch the basal learning process to execute their workflow efficaciously pinch AI, really tin you expect them to besides attraction connected the checks and balances of AI’s output without the due support from activity to return clip to create and refine wholly caller workflows?” she said. 
• “We’re focusing connected the surface-level reappraisal of AI’s output because that’s each group person the capacity and due support for. That information constituent is simply a immense activity SOS.”

Her halfway statement present aligns pinch the astir pointed statement successful her presentation: hole the infrastructure earlier you standard the output. AI won’t termination credibility. Untrained oversight will.

What smaller brands tin really do

The closing statement of Libert’s SMX talk, that AI rewards marque equity alternatively than creating it, raises a adjacent interest for newer entrants who deficiency years of accumulated authority. I pushed her connected this.

Her reply was much optimistic than the header suggests. AI Overviews aren’t simply surfacing the apical hunt results. They are surfacing brands that person built genuine authority astir niche topics. 

• “Younger brands tin still compete by focusing connected building retired their entity authority astir the agelong tail, wherever astir conversions really live,” she said.

She besides sees an opening created by the aforesaid firm inertia that slows large-brand adaptation. 

• “Plenty of larger brands are stuck successful much firm bureaucracy and are slower to accommodate to the changes and opportunities of utilizing AI to standard insightful information study and thought activity content. There’s really plentifulness of room for smaller brands to compete, now much than ever.” 

The caveat?

• “It’s a matter of slowing down and learning really to usage AI effectively, not scaling AI slop.”

The playbook for AI visibility

The 2026 AI hunt playbook that Libert presented astatine SMX distills to 4 imperatives: 

• Monitor marque practice crossed each influential platforms.
• Build entity authority done original investigation and taxable matter expertise.
• Triangulate visibility crossed search, video, societal proof, and trusted media.
• Govern AI usage pinch general disclosure and reappraisal processes alternatively than advertisement hoc workarounds.

None of this is peculiarly analyzable successful concept. The trouble is organizational. 

The brands that dainty credibility arsenic infrastructure alternatively than aesthetics are astir apt to beryllium cited, recommended, and trusted arsenic AI hunt continues to mature.