For most of my career, the defining promise of technology has been actionable information. Personal computers have made knowledge accessible, the Internet has made knowledge searchable, and mobile has put knowledge in our pockets. This change has reshaped commerce by making it easier for consumers to discover, compare, and transact.
New problems also arose. As access to information expands, decision-making becomes a bottleneck. Anyone who has ever turned to Dr. Google for a diagnosis knows that feeling. You won’t go home with a clear feeling. We are left with a long list of possibilities and little guidance about which possibilities are important.
AI relies on readily available information expertise Guidance, interpretation, and context are at your fingertips to help you decide what to do next.
This is not limited to healthcare or shopping. It will appear in finance, travel, education, customer support, corporate procurement, and any other area where there is a decision point between consumers and producers.
of consumer producer the system is about to change
At the highest level, commerce is an exchange between producing and consuming entities. In some cases, individuals may purchase the product. They may also be patients from whom businesses purchase services or seek care. The “what” is different, but the structure is the same. Something is needed, something is provided, and someone has to decide and take action.
Historically, the bridge between the two countries has been information. Search engines, marketplaces, comparison tools, and recommendation feeds helped buyers see their options. But they are information intermediaries. They surface options. They do not carry the burden of reasoning. That task is left to humans.
In an agent-mediated economy, that bridge is expertise, with agents on both sides. Agents on the consumer side represent intent, context, and constraints. Provider-side agents represent services, policies, capacity, and performance. These agents tailor the set of viable options to help consumers make the final decision.
The difference in medicine is obvious. Today, when you search for symptoms online, you are flooded with information rather than expertise. Soon, consumer health practitioners will be asking follow-up questions, eliciting relevant medical history (with permission), and inferring patterns just as clinicians would. Provider-side agents match that context against comparable cases, available paths, and constraints such as scope and timing.
Humans still make the decisions, but agents handle the information intake and reasoning that determines where the patient goes next. Instead of patients going to their GP with a vague complaint, they arrive with a clearer hypothesis and a recommended route to the right level of care.
why expertise, do not have personalization, is groundbreaking
It’s tempting to describe this change as personalization, as agents begin to understand your preferences, history, and constraints in a way that’s tailored to you.
But personalization is a result. The breakthrough is scalable expertise. These systems can interpret context, ask clear questions, and apply domain knowledge to guide decision-making. That’s the difference between systems that help you browse and systems that help you choose.
The job of the producer will also change. Once the expertise moves to the interaction layer, the producer must start designing for the agent’s interpretation. That means making it easier for agents to understand, verify, and trust what you offer.
As the agent sorts, compares, and makes initial inferences, the center of gravity shifts.
Buyers spend less time choosing options and more time approving results. Producers spend less time optimizing for attention and more time optimizing for reliable execution. And the systems underlying commerce are evolving from being designed for human viewing to being designed for agent coordination.
A new tension: the principal-agent gap
In the information age, the principal and agent were the same entity: the consumer. You searched, interpreted, and chose. There was no distance between intention and action. In the age of agents, that changes. Principals, whether they are the people making the purchases or the leaders making business decisions, set goals and constraints. The AI agent then processes the evaluation loop, narrowing down options, weighing tradeoffs, and recommending a course of action.
This change creates speed and clarity, but it also creates new gaps. Once the agent performs the first-stage decision-making task, the risk shifts from “did I find the right option?” “Did I choose the right tradeoff?” How do you make sure you understand what matters most? And how do you fix it if your agent optimizes towards the wrong outcome?
You don’t have to resolve that tension right away. But it will shape trust, market behavior, and the way producers design products in a world where buyers are increasingly agentic.
What does it mean for business leaders?
Once expertise becomes the new interface, the goal is simple for producers. It’s about being a business that agents understand, trust, and recommend.
It starts with making your products easy to read, including clean data, clear policies, and machine-readable structures. Agents also need to strengthen the relationship between what they promise and what they deliver, as they assess the gap faster than humans.
Conversely, if your business is consumer, make sure your buying agent has access to as much context as possible about what “good” looks like in terms of important parameters such as price, speed, reliability, and results. This allows the purchasing agent to do the comparison work and allows the team to focus on making trade-offs.
Companies that are ahead of the curve won’t treat this like a shift in marketing or purchasing. They treat this like an operational shift. This means you can design evaluations for your agents and then use them to drive faster decision-making around better judgment.
To learn more about AI and the future of work, subscribe to this newsletter.
