AI: 84 findings, with sources

Each AI assistant has its own source diet, so one visibility strategy cannot serve all three

The three assistants lean on visibly different source types. Google AI Overviews draws heavily on YouTube, Reddit and Quora. ChatGPT prefers publishers and news wires such as Reuters and the Associated Press. Perplexity tends toward niche and regional sites in health, finance and localised blogs.

The only sources cited by all three assistants are Wikipedia, Google, Amazon, Apple, Microsoft and two Brazilian portals

The seven domains that appear in the top 50 for all three assistants are named on the Venn: Microsoft, Wikipedia, Amazon, Google, Apple, Globo and UOL. Five global giants plus two Brazilian portals - meaning the only universally cited sources are the largest reference properties on the web, not the sites most brands can realistically join.

Magalu's AI-powered virtual influencer Lu reached 4.5 million WhatsApp users in five months, generating R$28 million in sales at 3x the conversion rate of Magalu's own app.

Magalu gave its 23-year-old virtual influencer Lu an agentic AI brain on WhatsApp: 4.5 million users in five months, R$28 million in sales, conversion 3x higher than Magalu's own app. A beloved brand asset can become the AI interface itself, turning agentic commerce from cold automation into cultural connection.

Dove's campaign publishing unedited Reddit reviews drove sales up 108.4% year-over-year and made it the most ChatGPT-recommended hair mask, appearing 58% of the time and being recommended 89% of the time.

Dove invited Reddit's brutally honest community to review its hair mask, publishing the first 50 reviews unedited as the campaign: sales +108.4% YoY, #1 presence on ChatGPT (58%), recommended 89% of the time. Authentic community conversations feed AI visibility.

Human and AI effectiveness scores of 480 Cannes ads show zero correlation.

480 Cannes-entered ads scored on effectiveness by humans and by Gemini show zero correlation - 'the regression line is totally flat.' AI does not judge creativity like a human and must not be treated as a surrogate human judge.

Structured ad metadata explains AI's judgments about 3x better than it explains humans' judgments.

Structured metadata parameters (~30: duration, product presence, narrative, celebrity...) explain AI ad judgments about 3x better than human judgments - AI evaluates advertising systematically where humans don't.

AI and human audiences rank almost opposite creative traits as their top two preferences.

AI's top two creative preferences: ads that clearly feature the product, and real people rather than celebrities. Humans' top two: emotional intensity and narrative. The two audiences reward nearly opposite creative choices.

Long-term brand equity drives 63% of LLM visibility.

63% of LLM visibility is driven by long-term brand equity (Charlie Oscar: 30 brands, 6 categories, 8 months). You can't buy reach across LLMs - they reflect back distinctive assets, consistent positioning and long-term branding.

A quarter of Americans use AI daily; intensity is growing faster than adoption

25% of Americans now use AI daily, up from 19% in 2025; among AI users, daily usage jumped from 31% to 39%, and 24% spend more than an hour a day with AI. Overall US adult adoption moved only from 61% to 64% - intensity is growing much faster than reach.

Paying for AI flipped from 3% of users to 55% in one year

55% of AI users now pay for at least one AI product, up from 3% in 2025. The typical payer spends USD 20-49 a month; 48% cover the cost entirely themselves, 34% have family or friends paying, and 20% are covered by an employer or school.

For medical, legal and mental-health advice, humans beat AI three to one

For high-stakes advice, humans still win by roughly 3 to 1: 63% prefer a human for medical diagnosis (19% AI), 61% for mental health support (19% AI), 59% for legal advice (21% AI). Even for learning new skills, humans lead 44% to 28%.

NatWest converted 10,000 real consumer voices into schema-tagged data seeded across AI platforms, lifting AI-visible answers by 508% and driving 500,000 new savings accounts, up 168% year-over-year.

NatWest captured 10,000 real consumer voices about savings, transformed them into schema-tagged semantic datasets and seeded them across high AI-authority platforms: +508% growth in AI-visible answers, #1 bank for machine-cited savings advice, and 500k new savings accounts (+168% YoY).

The US will lead global agentic AI consumer spending in 2030 with 31.9% ($1.1 trillion), ahead of China (15.1%) and the UK (4.9%).

The US will account for 31.9% of global agentic AI consumer spending in 2030 ($1.1 trillion), followed by China (15.1%, $505.8bn) and the UK (4.9%, $131.2bn).

In habit-driven categories, brands should win the "default reorder" - getting consumers to tell an agent to order a specific brand rather than a generic category - through consumer-facing ads and cultural relevance that the agent then executes.

In habit-driven categories, the brand's job is to win the default reorder: make consumers tell an agent 'order Pepsi' rather than 'buy a case of cola'. Consumer-facing ads and cultural tie-ins build the preference the agent then executes.

One in four sources AI models cite is a YouTube video.

One in four sources cited by AI models is a YouTube video (Share of Model data). Models seek consumer consensus - reviews, recommendations, how others talk about you - so activating your brand's echo matters as much as your own messaging.

Daily AI assistant use beats ride-sharing and food delivery combined

General AI assistants are used daily by 22% of US consumers - more than ride-sharing (6%) and food delivery (7%) combined. In two years, AI assistants became a more frequent habit than the defining consumer apps of the last decade.

A third of agent users have let AI act without final approval

41% of AI users have tried AI agents, 24% use them regularly, and 32% have already let an AI act on their behalf without final approval. Delegation - not just conversation - is becoming a consumer behaviour.

92% of AI agent users pay for AI - agents are the monetisation frontier

Agents and money travel together: 92% of AI agent users pay for AI, 63% use AI daily, and 65% of AI payers use agents versus 13% of non-payers. Agentic capability is where consumer willingness to pay concentrates.

Accuracy and trust now beat ease of use in choosing AI tools

Consumers pick AI tools on substance, not slickness: accuracy (45%), trustworthiness (40%) and security/privacy (36%) lead the selection criteria, while ease of use fell from 38% to 32% year over year. Table stakes moved from UX to reliability.

Known AI-generated content carries a 2-to-1 engagement penalty

36% of consumers say they are less likely to engage with content they know is AI-generated; only 14% are more interested. For brands leaning on visibly AI-made creative, the label itself carries an engagement penalty of more than 2 to 1.

Nearly half of AI users already talk to their AI; legacy assistants are fading

Voice is becoming a genuine AI interface: 46% of AI users speak to their AI at least sometimes and 19% use voice as their primary mode - while the legacy voice assistants slide (Alexa down from 32% to 22% reach, Siri from 25% to 20%). Menlo's 2027 prediction: voice moves from dictation to instruction.

AI now assists the majority of writing, coding and assignment work

Where AI penetrated everyday tasks in 2026: 61% of people who write use AI for writing support, 58% for coding projects, 56% for school or work assignments, 53% for image creation - while among all US adults, researching topics (26%), staying in touch (25%), healthcare navigation and paying bills (23% each) show AI seeping into life admin.

Agentic AI-mediated consumer spending will triple from $944 billion in 2026 to $3.35 trillion by 2030.

Agentic AI will facilitate $944 billion of global consumer spending in 2026 (1.3% of private consumption), tripling to $3.35 trillion by 2030 (3.8%).

Telecoms & utilities will grow 611.9% to become the largest agentic commerce category by 2030, at $410.3 billion.

Telecoms & utilities will grow 611.9% to become the single biggest agentic commerce category by 2030 ($410.3bn) - information-dense decisions, frequent billing and comparison-led switching make it ideal for delegation to AI agents.

A purchase journey doesn't need to be low-value to be delegated to an AI agent - it needs to be repetitive, searchable, and measurable.

A purchase journey doesn't need to be low-value to be delegated to an AI agent - it needs to be repetitive, searchable and measurable.

Soft drinks will see the fastest growth in agentic AI purchasing by 2030 at +403.5% ($304.8bn), ahead of media & publishing (+401.8%, $367.8bn) and food (+274.8%, $292.8bn).

Soft drinks will see 403.5% growth in agentic AI purchasing by 2030 ($304.8bn) as habit-led, low-value replenishment becomes easy to automate; food grows 274.8% ($292.8bn); media & publishing grows 401.8% ($367.8bn).

In high-value or sensitive categories like automotive, pharma, and electronics, AI agents gain influence over the shortlist but people retain the final purchase decision, with only modest, mostly horizontal movement on the Four Modes map.

In high-value or sensitive categories (automotive, pharma, electronics), agents gain influence over the shortlist but people keep the final say - movement is modest and mostly horizontal in the Four Modes map.

The top 14% of AI spenders account for 60% of all consumer AI spend

The top 14% of consumer AI spenders - those paying USD 100+ a month - account for 60% of all consumer AI spending. Power spenders are 3x more likely to earn USD 100k+ and 2.5x more likely to be regular agent users.

53% of marketers value AI more for growth than efficiency, and 57% of CMOs rank effectiveness as AI's most important future value.

53% of marketers value AI more as a growth driver than an efficiency driver, yet believe their C-suite sees AI purely through a productivity lens (McKinsey); 57% of CMOs rank effectiveness as AI's most important future value - the only C-suite group to do so (BCG/NMA).

Consumer AI spending tripled to USD 40 billion while users grew just 11%

Global consumer AI spending reached USD 40 billion in 2026, more than triple the USD 12 billion of 2025 - while global AI users grew only 11%, from 1.8 to 2.0 billion. The market's story shifted from adoption to monetisation in a single year.

The average AI user now runs three assistants side by side

AI users stack assistants instead of choosing one: the average user now runs 3.0 general AI assistants, up from 2.2 a year ago. ChatGPT reaches 60% of AI users (45% in 2025), Gemini 58% (36%), and Claude nearly tripled to 20% (from 7%).

Nearly a third of US adults have already earned money with AI

48% of AI users - roughly 30% of all US adults - have earned money with AI, led by content creation (22%) and freelancing (18%). The generational gap is stark: 62% of Gen Z AI users have monetised AI versus 15% of Boomers.

Distrust of AI-generated information among non-users rose from 58% to 70%

AI holdouts are hardening, not softening: 83% prefer dealing with humans, 76% cite privacy concerns, 52% believe AI tools are biased, and distrust of AI-generated information rose from 58% to 70% in a year. The remaining third of non-users will not be won by feature lists.

38% say AI improved their lives; half feel no change at all

Consumer sentiment about AI is calm rather than euphoric: 38% say AI has improved their lives, 52% report no meaningful change, and 10% say it made life worse. Meanwhile 43% of working adults see moderate-to-significant risk to their job within five years (46% of Gen Z).

Parents and higher-income households lead AI adoption by 20+ points

AI adoption maps onto money and parenthood: 77% of US adults in households earning USD 100k+ use AI versus 56% under USD 50k, and 84% of parents use AI versus 56% of non-parents - with 40% of parents using it daily.