Agentic Commerce Part 2: How to prepare your store for AI shopping

Are you ready for AI that shops for your customers?

In Part 1, we explored why Agentic Commerce might not be just another AI buzzword—and how signals from Amazon, Adobe, Shopify, and others point to a real shift in how digital commerce is evolving.

So what should retailers actually do to prepare?

Let’s start with the foundation: communication between systems.

Enter the Model Context Protocol (MCP)

To succeed in an Agent-First world, retailers need more than a traditional website. They need machine-readable ecosystems that AI agents can understand and transact with.

This is where Anthropic's MCP (Model Context Protocol) comes into play. It is a developing standard that provides tools and context to AI Agents, enabling brands to communicate product, pricing, and availability data in formats that AI agents can read and act on.

In simple terms, MCP helps your backend systems communicate effectively with AI Agents. MCP along with structured data, becomes the key to enabling machine-to-machine transactions and product discovery in Agent-Led journeys.

The next step Agent-to-Agent Commerce

The next step will be for those agents to negotiate and complete transactions directly with other Agents. That’s where  where Google's new open protocol Agent-to-Agent (A2A) comes in. Developed with over 50 technology partners, A2A enables AI agents to communicate, share information, and coordinate actions across different platforms.

For retailers, that means your own AI Agents will be able to speak directly with other Agents—negotiating prices, checking stock, and completing purchases—without human involvement. Complementing MCP, A2A provides a universal way to connect, manage, and integrate agents from different providers, paving the way for faster, smarter, and more automated commerce.

So what does this mean for Online Retail?

Whether you're fully convinced or still cautious about Agentic Commerce, the implications for online retail are already becoming visible. Here is what is changing and why it matters.

1. Visibility and understanding of AI Traffic becomes critical

Not all AI traffic is the same. Some agents scrape your content while others generate real sales opportunities. Retailers will need to differentiate between profitable agent traffic and performance-draining bots, then optimise accordingly.

2. Discovery shifts to Agent-Led journeys

As AI tools begin to replace traditional search engines, product discovery will be driven by large language models rather than search engine results pages. Retailers must begin to structure their data for machines as well as for people.

MCP and structured data will become the new front doors to your storefront, enabling AI agents to interpret product information, availability, and pricing in real time.

3. Commerce moves beyond websites

In the near future, customers may not visit your homepage at all. Expect more transactions to occur through Agent-to-Agent (A2A) interactions, where your backend negotiates pricing, stock, and delivery with other systems.

Your site will still matter for brand and human visitors, but much of the actual commerce could shift into backend exchanges between agents, powered by MCP and structured data.

4. AI Agents raise the bar for performance

While AI agents are patient, they optimise aggressively. A slow site or a broken checkout process could result in your offer being ignored or downranked in an agent’s internal selection process.

This makes Load Testing and Real-user Journey Monitoring essential for maintaining performance under new conditions.

Customer journeys now span across devices, bots, assistants, and digital backends. It will no longer be enough to monitor how a human travels through your site. You’ll soon need to monitor how your systems perform for both human and machine visitors alike.

Where to begin (realistically) with Agentic Commerce

This shift may not require immediate action from every retailer. You don’t need to overhaul your operation overnight. But ignoring this disruptor could lead to missed opportunities and reduced visibility as the landscape changes.

Here are seven practical steps to get ready for AI shopping:

1. Prepare your backend for MCP compatibility

2. Use monitoring and analytics tools to understand AI agent traffic and conversions

3. Identify valuable AI agents and distinguish from scraping bots

4. Reframe your SEO strategy to prioritise AI-based discoverability

5. Ensure product data is accurate, structured, and machine-readable

6. Conduct load testing for increased machine-to-machine transaction capacity

7. Consider developing your own AI agents to operate in this growing ecosystem

Final thoughts

Agentic Commerce is not just a buzzword or a new channel. It represents a fundamental shift in how eCommerce operates.

Retailers who adapt to serve both humans and machines will unlock new revenue opportunities, reduce friction, and future-proof their customer experience—even when their customers are not the ones doing the shopping.

The real question is not whether Agentic Commerce is coming. It is this: How ready will you be when it arrives?

Look out for our downloadable Checklist a complete 7-step preparation guide coming soon.

Further resources

Read Part 1 of our blog 'Agentic Commerce Part 1: Hype or the next big eCommerce shift?'

If you'd like to learn how monitoring and Load Testing can help you prepare for AI-powered commerce, get in touch.