AI is about to make the buying journey more efficient. That does not necessarily make it more trustworthy.
In a conversation with Adweek’s The Speed of Culture podcast, Human Ventures Holdings CEO Joe Marchese argues that marketers are entering a world where AI agents could influence purchases before a brand gets a direct chance to speak. The important shift is not simply that machines will recommend products. It is that the recommendation may become the first serious brand interaction.
That changes what attention is worth, what brand means, and where marketers should spend their effort.
AI agents turn attention into a trust problem
Marchese’s argument starts with an older lesson from his time at true[X], the advertising company he founded before its acquisition by Fox. As Adweek summarizes it, data has limited value if a brand never earns a person’s attention. An impression may be measurable, but measurement does not make it meaningful.
AI agents introduce a harder version of the same problem. If an agent acts on a consumer’s behalf, the consumer needs to believe that the agent is working for them. Marchese compares that relationship with human agents in areas such as real estate, insurance, and entertainment. The expectation is not just assistance. It is aligned incentives.
Advertising complicates that promise. An agent that recommends the product paying the most rather than the product that best fits the user may still look helpful on the surface. Underneath, it is closer to hidden persuasion. The interface becomes conversational, but the incentive problem remains familiar.
That is why the next battle for brands may not be about appearing inside an AI answer. It may be about becoming the brand an agent can safely recommend. Product quality, reputation, clear evidence, and a coherent point of view all become more important when the consumer is no longer browsing a shelf of competing messages.
The advantage moves toward what machines cannot flatten
Marchese’s second argument is more uncomfortable for performance marketing. As Adweek reports, he believes anything machines can understand will eventually be optimized. If every competitor can automate targeting, testing, bidding, and variation, those capabilities become table stakes rather than durable advantages.
The remaining advantage sits in areas that are harder to reduce to a pattern: strategy, culture, experiential marketing, and real-world human connection. That does not mean brands should abandon data or automation. It means optimization cannot be the whole strategy.
Marchese points to live events, festivals, hospitality, and other shared experiences as examples of places where brands can create value that is difficult to reproduce through a personalized interface. A physical experience gives people something to remember together, not just something an algorithm selected for them.
This also explains why brand cannot be reduced to the person most likely to click. In the discussion, Marchese rejects the idea that everyone outside a narrow target audience is wasted reach. A brand also exists in the conversations around a buyer, including friends, family, colleagues, and wider culture.
That broader role matters when AI agents become more influential. A machine may be able to compare specifications and prices. It will be less capable of understanding why a brand carries meaning in a community, why a live moment becomes culturally important, or why people want to be associated with something before they are ready to buy it.
The strategic consequence is clear. Brands should prepare for AI-mediated discovery by making their products easy to evaluate and their claims easy to trust. But they should not mistake machine-readable evidence for cultural relevance. As buying decisions move closer to automated recommendation, the brands with the strongest position may be the ones that combine both: enough clarity for the agent, and enough meaning for the human who ultimately lives with the choice.