Why it matters
AI discovery is moving into everyday behaviour, but usage is not the same as unquestioned trust. New IAB research shows frequent AI use alongside routine verification, while publishers are still negotiating how their work is attributed, measured and paid for. For marketers, that creates a practical shift: visibility inside an answer is useful only when the underlying claim can survive the next click, comparison or source check.
01 · Read both signals
Adoption is strong. Trust is conditional.
On 16 July 2026, IAB released two studies on consumer AI use and AI-driven content discovery. The consumer study surveyed 500 US adults who had used an AI tool in the previous month. Among those users, 72% said they use AI at least weekly, 40% reported daily use and more than 80% described their overall experience as positive.
The trust indicators are more revealing for marketers. Fifty-seven percent said they often or always verify information from AI, while 83% said it is important that AI tools explain how their answers are generated. More than 60% said the reputation of the company behind a tool affects their trust. People are not waiting for perfect confidence before using AI; they are building verification into the behaviour.
The boundary matters. This is a self-reported US survey of recent AI users, not a global measure of the whole population, and it does not prove how every person verifies a commercial answer. My inference is narrower: discovery can scale faster than confidence. A brand may earn visibility inside an AI response and still lose the decision when someone checks the source, date or evidence behind it.
02 · Follow the value exchange
Discovery is becoming a commercial negotiation.
IAB's companion study surveyed 110 US publishers with annual revenue above $100,000 whose respondents were familiar with their company's LLM strategy. Ninety-five percent reported that LLM-driven discovery is having a moderate or significant effect on their organisation, and 80% described the net impact as positive. That optimism comes with conditions.
Fifty-one percent of the publishers surveyed had signed at least one AI or LLM licensing agreement and another 35% were negotiating. The distribution is uneven: IAB reports signed agreements among 60% of larger publishers but only 20% of small independent publishers. Publishers also prioritised referral traffic, downstream measurement and revenue sharing—not payment alone.
For marketers, this is not only a publisher-industry issue. AI answers depend on an evidence supply chain. If credible sources restrict access, change licensing terms or invest more heavily in owned audiences, the material available to support a brand claim changes too. My view is that content strategy now needs a sourcing strategy: which claims depend on third-party authority, which evidence the brand owns, and whether each source remains accessible and attributable.
03 · Design for verification
A visible answer can still fail the source check.
The wider market is already moving towards more explicit provenance. On 9 July, Google announced a global 'How this ad was made' section in My Ad Center for ads across Search, YouTube and Discover. Google says ads created with its own generative tools will receive an automatic disclosure, while advertisers can indicate when external generative tools were used. Depending on local requirements, a label may also appear directly on the ad.
That is transparency infrastructure, not a guarantee of truth or compliance. Google's policy notice explicitly says its AI label setting does not guarantee compliance with specific regulations. C2PA standards can certify the source and history of media content, which is useful provenance, but provenance does not validate the commercial claim inside an asset. Marketers still need substantiation, rights review and accountable approval.
My inference is that traceability is becoming part of performance. The useful questions extend beyond 'Was the brand mentioned?' to 'Was it cited accurately?', 'Could a person reach the original evidence?', 'Was the source current?' and 'Did the interaction return value to an owned channel?' Those are not yet standard dashboard fields, but they are closer to decision quality than a raw count of AI mentions.
04 · Change the operating model
Build the trust layer before buying more visibility.
This is not a request for every article to become a research paper. It is a request for high-consequence claims to have an owner and an evidence trail. A product capability, market statistic, performance result or regulatory statement should connect to a current primary source, a clear date, the right usage permission and a review record that someone can reconstruct.
The same discipline should carry into distribution. Canonical URLs, structured metadata, descriptive titles and stable source links make content easier for people and machines to interpret. But technical hygiene cannot rescue weak evidence. The sequence matters: substantiate the claim, publish it clearly, preserve provenance, then optimise how it is discovered.
I would also separate AI visibility reporting from business acceptance. Monitor whether priority questions surface the brand and cite the right page, but judge the channel through assisted research, qualified visits, direct audience growth and commercial outcomes where measurement allows. The goal is not to be quoted everywhere. It is to become a source that a buyer, publisher or AI system can use without lowering the quality of the decision.
Reach inside an answer is only the first step. Trust is earned when the evidence remains intact after the answer is challenged.
What marketers should do next
Turn the signal into a better decision.
- Create a claim register for the statistics, product capabilities, customer results and regulatory statements most likely to influence a decision.
- Attach a current primary source, publication date, rights status and accountable owner to every high-consequence claim.
- Test priority buyer questions in major AI discovery experiences and record whether the brand appears, which page is cited and whether the citation is accurate.
- Measure AI visibility separately from acceptance signals such as qualified visits, return behaviour, owned-audience growth and assisted commercial outcomes.
- Review publisher and platform terms so licensing, attribution, content access and downstream measurement are explicit rather than assumed.
Sources & further reading
01IAB — The Next Phase of AI: From Adoption to Accountability02IAB — Consumer and publisher research release and methodology03Google Ads & Commerce — Expanding AI transparency in ads04Google Advertising Policies — Updates to AI labelling requirements05C2PA — Technical specifications for content provenance