Your next customer might be an AI agent: what agentic buying means for your business

Agentic buying is when a buyer hands part of a purchase to an AI agent: the agent finds suppliers, compares products, prices and delivery terms, builds the cart or the quote request, and completes the order once the buyer approves. Until recently this was a demo. On October 13, 2026, it became infrastructure. Shopify launched Shop MCP, a single connection that lets any compatible AI agent search more than one billion products, build carts and check out with Shop Pay, no partnership agreement required. On the business-to-business side, Gartner predicts that by 2028, 90 percent of B2B buying will be AI agent intermediated. For a traditional company that sells to other businesses, the practical meaning is simple. Your next buyer may reach you through software first, and that software will not wait for a callback.

What changed this month

Shopify's move matters because of how open it is. Previously, an AI agent could find a product but usually had to send the shopper to the store's website to pay, and many buyers dropped off along the way. Now, according to the Shopify developer changelog, an agent can search the catalog, build a cart and call checkout at the store that sells the item, completing the purchase after the shopper confirms the total and the payment method. Search works without credentials, so any agent builder can plug in. In addition, the launch runs on MCP, the open standard for connecting AI to business systems that we covered in MCP just became a real standard. In other words, agent shopping is no longer a set of private deals between large platforms. It is becoming a public road that any assistant can drive on, and that changes who your storefront is really built for.

Why this reaches far beyond online stores

It is tempting to file this under retail and move on. However, the bigger shift is in business buying, where purchases are larger and more repetitive. Gartner's prediction that 90 percent of B2B buying will be agent intermediated by 2028 comes with a second warning: products will need to be machine-readable, and traditional search optimization will give way to what Gartner calls agent engine optimization. Meanwhile, buyers are already halfway there. Forrester's 2026 State of Business Buying research reports that nearly all business buyers, 94 percent, use AI during their buying process. Consider what that looks like at a metals distributor or a packaging supplier. A procurement manager asks an assistant to find three suppliers who can deliver a part within a week. The assistant reads websites, catalogs and price sheets, then shortlists the ones it can understand. A supplier with a vague website and a PDF price list from last year simply does not appear.

Why this is an opening for a smaller business

Agent-led buying sounds like a game for large companies with big e-commerce teams. In practice, it can favor the smaller seller. An agent does not care about brand budgets or how many salespeople you employ. Instead, it rewards three plain things: accurate information, clear terms and a fast answer. A 30-person distributor can win on all three, often faster than a large competitor with layers of approval. We have seen the same dynamic in sales without agents: in distribution, the first quote back usually wins the order. When the buyer is an agent comparing five suppliers in seconds, that advantage gets sharper. Consequently, the question is not whether your business needs a shopping agent of its own. It is whether your catalog, pricing and order process are ready to be read and answered by somebody else's agent.

What "agent-ready" means in practice

Being ready does not require new software on day one. First, your product and service information has to exist in one place, with specifications, units, lead times and minimum quantities written down rather than living in a senior salesperson's memory. Second, that information has to be current, because an agent that finds a wrong price will either skip you or create a dispute. Third, the path to ordering has to be clear: how to request a quote, what the terms are, who approves exceptions. Fourth, your response time has to be measured in minutes, not days. This is where automation creates direct value for a business without an R&D department. An agent that drafts quotes from your price list for a person to approve, or one that keeps stock and catalog data in sync, is a modest project. Our guide to back-office automation with AI agents covers the building blocks.

What to do this quarter

1. Ask an AI assistant about your own products. Pose the question a buyer would ask, such as "who supplies X with delivery in a week," and see whether you appear and whether the answer is correct.
2. Put your catalog in one source of truth. Specs, prices, lead times and minimum orders in one system, kept current, then published clearly on your website.
3. Cut quote turnaround first. Automate the draft, keep a person on the approval. This pays off today with human buyers and tomorrow with agent buyers.
4. Write your rules for agent orders. Decide which order sizes or repeat purchases can be confirmed automatically and which always need a person on your side, the same human-approval principle we apply in AI agents for small businesses.
5. Watch the security side. Treat incoming agent traffic like any new channel, with clear limits on what it can access, as we describe in AI agent security.

Frequently asked questions

What is agentic buying?

It is when a buyer delegates part of a purchase to an AI agent that searches, compares, builds a cart or quote request and completes the order after approval. The buyer still decides, but the seller's first contact is often with software.

What did Shopify launch in October 2026?

Shop MCP, launched October 13, 2026, lets any MCP-compatible agent search more than one billion products in Shopify Catalog, build carts and complete Shop Pay checkout after the shopper confirms the total and payment method.

Does agentic buying matter for B2B companies?

Yes. Gartner predicts 90 percent of B2B buying will be agent intermediated by 2028, and Forrester found 94 percent of business buyers already use AI while buying. Sellers that agents cannot read quickly risk missing the shortlist.

How can a small business prepare?

Start with information: one current source for products, prices and lead times, published clearly. Then speed up quoting with an agent that drafts for a person to approve, and set rules for which agent orders need a human. Our implementation guide shows how to pick the first process.

References

• Shopify developer changelog: Add shopping and Shop Pay checkout to any agent with Shop MCP (October 13, 2026)
• Gartner: Gartner Unveils Top Predictions for IT Organizations and Users in 2026 and Beyond (October 21, 2025)
• Forrester: The State Of Business Buying, 2026 (January 21, 2026)
• In distribution, the first quote back usually wins the order
• MCP just became a real standard