A quiet switch just flipped in the economy. For the first time, large language models can do consumers’ shopping for them. You can now ask ChatGPT to recommend a product and—thanks to a new Instant Checkout feature—buy it right in the conversation, without ever leaving OpenAI’s website. Etsy and Shopify stores are live in ChatGPT; Walmart is on deck.
Under the hood, the new Agentic Commerce Protocol (ACP) lets AI agents and merchants speak a common language to pass inventory, terms, and payment tokens back and forth. The interface you use is still early and it can be glitchy, but the underlying technology is rolling out.
Soon enough, many of us may shop through an AI chatbot rather than on a traditional retail website. “I want the best non-stick frying pan under $150.” Boom—done.
This development is being touted as “agentic commerce” or “intelligent commerce.” And it’s a crossroads moment.
If agentic commerce is built the way the internet works today, we’ll get an economy even more by advertising and other perverse incentives—one where your buying decisions are based on obscure deals between online retailers and AI companies, where all the intimate conversations you share flow into an ad platform, and where manipulative AI chatbots whisper in your ear over the months to spend more, or more foolishly.
It doesn’t have to be that way. If we’re intentional, we can use this moment to build something much more consumer-centric—an economy where consumers have more power than ever before. This is what Consumer Reports (CR) and Stanford Digital Economy Lab, along with industry partners, are organizing to deliver.
Introducing Conscious Commerce
Imagine an AI that helps people shop in line with their values and goals—and is loyal to the user’s interest, not the platform’s. Instead of “agentic commerce,” let’s call this “conscious commerce.”
Conscious commerce is agentic commerce that knows what you care about, and is loyal to you. It’s the fusion of personal intelligence and consumer alignment: systems that don’t just predict what you’ll click, but help you act in integrity with your preferences and values.
Let’s take a major decision—replacing a busted appliance, deciding on a used car, or comparing child‑care services. Instead of spawning 50 tabs and stressing the details, you instruct a digital agent to sift through the information for you. It understands your priorities—factors like budget, noise level, repairability, safety ratings, and budget. It requests similarly structured offers from sellers, and highlights the trade‑offs that matter specifically to you.
Those might include the long-term cost of ownership (taking repair rates into consideration), or the company’s labor practices and carbon emissions. If a seller pays a referral bounty to a site offering the product you want, your agent flags that as a conflict and asks if it should rerank the options. You can ask the AI, “Why this recommendation?” and see the reasoning behind it. Then you approve a one‑time payment token. No shadow profiles or data sprawl created.
Conscious commerce is AI‑mediated commerce that centers individual, human intentions and values—not a platform’s or seller’s agenda. An agent represents you, carries your now-portable preferences wherever you shop, and operates under a duty of loyalty to you, transparently revealing any conflicts of interest. Think of it as moving the whole internet model from “clicks and ads” to intent and alignment.
Far-seeing technologists like Doc Searls have long argued that we need to move from an “attention economy” to an “intention economy.” With advances in agentic AI, it could finally happen. But for this to work, consumers need certain guarantees about the agents they’re being asked to trust. This might mean, for instance, agents that are contractually or legally obligated to adhere to ethical standards that are much stricter than what we’re used to online—more like the professional discretion required of doctors and lawyers.
Designing Standards—Now
Every major platform is moving toward agentic shopping interfaces. OpenAI, Google, Amazon, and others are all building transactional agents that learn from context, try to anticipate your needs, and complete purchases on their own. The way these efforts play out could define the next century of commerce.
The infrastructure for conscious commerce won’t build itself. Right now, CR and Stanford’s Digital Economy Lab are inviting partners—retailers, payment networks, AI labs, policymakers—to help design that architecture together. We’re already running multi‑stakeholder workshops and technical sprints with companies across the stack to turn these ideas into open standards and reference implementations.
We’re doing this work because we want to see tools that better serve their users. But we also think that tools like these could have a transformative effect on the economy, making a healthier marketplace for businesses as well as consumers.
New Power for Consumer Preference
We’ve already described a future marketplace where you can trust a digital agent to recommend products that align with your values. Now, scale that up.
Imagine millions of people setting preferences that are fully portable, applying across every consumer transaction. Preferences like “choose long-lasting goods”, “avoid forced‑labor supply chains,” “buy local within 15 miles,” or “support accessible design.”
Build a system like that, and watch suppliers respond. They publish agent‑readable information (repairability, software‑update commitments, parts pricing). Retailers tune promotions to long‑term value, not impulse buys. Discovery shifts from ad impressions to verifiable attributes.
Preference gravity reshapes the aisle.
In aggregate, our individual choices can roll up to collective bargaining power for fairer, safer, and greener products. Because consumer power is one of the most powerful forms of human agency—it spans what people need, what they value, and what they buy.
We’ve been here before. In the 1930s, civic leaders feared the distorting effects of advertising. It was a time of low trust, when fact and puffery were sometimes indistinguishable. Colston Warne, an economist and one of the founders of Consumers Union, warned that “uninformed choice is not free choice.” The consumers’ movement was formed around a simple idea: Ordinary people deserve independent, objective information before they buy.
These challenges will take on epic dimensions of scale by the 2030s. In a marketplace exploding with choice and mediated by machines, how does an individual keep control? How can we make sure the algorithms that guide our spending are accountable not to advertisers, but to us?
These are questions for all of us, because we’re all consumers.
And they also pose immediate questions for businesses. Retailers, marketers, and payment companies are putting a lot of faith in agentic commerce. But shoppers will only delegate purchase decisions to autonomous agents if they feel confident in how their data is handled, and proof that AI is helping consumers rather than just pushing profitable products (How many people would trust an AI to go shopping with their credit card today?).
Trust—in how agents handle consumer data, and how they make recommendations—will be the gating factor for adoption. Shared rails for loyalty, transparency, and alignment can accelerate the development of agentic commerce and make the pie bigger for all sorts of businesses.
Open standards have always done that—from email to credit cards to the web itself. Each of these innovations were good for business and helped empower consumers. And every technological shift is an opportunity to reprioritize the public interest.
If we get conscious commerce right, agents will amplify what’s best in us as individuals—our discernment, our taste, our collective sense of value. If we don’t, we’ll wake up in an economy where consumer surveillance is an order of magnitude more intrusive than it already is, where consumer buying decisions are manipulated in ways we’ve never seen before, and where the goal of a fair and transparent marketplace has become impossibly remote.
Right now, people fret a lot about whether AI is aligned with human values. Do we really want a marketplace that anticipates every want before we have it, steers our buying behavior, and treats desire itself as just another thing to optimize? Who’s buying and who’s selling?
