AI agents could shift market power towards consumers, but the new battleground is trust in the agent's incentives
Companies have spent decades profiting from our inattention. AI agents could change that, but they will not necessarily make the market fairer.
Patrick Collison, Stripe's co-founder, recently described companies as the first superintelligence. The point is that companies can devote enormous resources to understanding and influencing our behaviour, while we have neither the time nor the information to respond.
Think hidden fees, friction designed to discourage cancellations, and paid product placements that influence what we buy. Entire business models depend on consumers not having the time, energy or information to make the best decision.
Personal AI agents could change the economics of that relationship. An agent working for the buyer can compare products, scrutinise terms, identify hidden costs and negotiate on the buyer's behalf. In theory, businesses would have to compete on product quality rather than their ability to manipulate the purchasing process.
Early evidence is encouraging, although far from conclusive. Linas's Newsletter reports that agents given the same brief booked the same Manhattan hotel in up to 100% of test sessions, while shoppers arriving through AI converted around 60% better in the retail data it cites. These are preliminary findings, not independently verified proof.
But there is a catch.
An agent does not eliminate the power imbalance. It creates a new one.
Who pays for the agent? Who determines what it optimises for? And what happens when the interests of the buyer conflict with those of the platform operating the agent?
There is a second problem. Agents can only compare what they can measure. Price, specifications and verifiable reviews are relatively straightforward. Durability, service quality and the value of a genuinely better experience are harder to assess. If an agent cannot recognise those differences, businesses may optimise for what the machine can measure rather than what the customer actually values.
The opportunity is enormous, but the outcome is not predetermined.
For decades, businesses have invested in exploiting the gap between what consumers know and what they have time to investigate. AI agents could narrow that gap, shifting competitive advantage towards better products and services.
The real battleground will be trust: can an agent independently verify a product's claims, and can the buyer trust the agent's incentives?
That is the question that will determine whether agentic commerce genuinely benefits consumers or simply creates a new intermediary with a different set of interests.
#AgenticAI #ConsumerTrust #AusBiz
Source: Linas's Newsletter, Patrick Collison: AI Agents Will Make Better Products Win (https://lnkd.in/gsezyZMb), 16 October 2026.