Is Your Commerce Stack Readable, Actionable, and Trustworthy for AI?

Change can be incremental, as in the cautionary “boiling frog” tale where a gradual rise in temperature leads to an unfortunate demise before the frog realizes what’s happening. Or change can be fast and furious; look at the disruption artificial intelligence is driving through the traditional path to purchase. Walmart’s introduction of an in-platform app experience in ChatGPT shows how fast things are moving. It’s led by its commerce agent Sparky, following the rollback of ChatGPT’s plans for transactions via its prototype Instant Checkout functionality.

This new integration highlights how discovery, conversion, and fulfilment are no longer separate steps on the shopper journey. The AI agents, from ChatGPT to our own Shopware Copilot and other external assistants, now collapse these stages into one direct conversation with customers. The agents will handle shopping, payments, and order management on behalf of the shopper and merchant.

Consumers are stepping away from beginning their shopping mission with a keyword-led search. Natural language is now coming to the fore. For instance, a traditional search for “cheap red running shoes sale” is likely to be replaced with a conversational ChatGPT query, such as “What are the best red running shoes under $80 for beginners?”

If you’re not on the list of recommended brands surfaced in the LLM answer box, you’re now 10 steps behind the frontrunners for a sneaker sale. Why? Because customers can seamlessly move from discovery in the LLM to purchase in the merchant’s owned environment. What used to take hours of searching and tab-hopping now happens in seconds.

The traditional ‘storefront’ or homepage is becoming just one of many endpoints in the transaction journey, rather than the primary touch point. Commerce agents will discover and transact in multiple environments, and your entire system needs to adapt to the new realities of fulfillment.

ChatGPT is signaling the shape of things to come. It is now focusing purely on discovery and has turbocharged its Agentic Commerce Protocol (ACP) to enhance the experience for AI assistants. The updated ACP improves the speed of response to an intent-driven shopper question and acts as the connective tissue with merchants. This means brands can integrate their preferred systems to present product data, images, promotions, and other relevant features via the technology to help clinch the sale.

What should you be reviewing regarding your systems? First, you need open, flexible infrastructure, so you can plug customized software via APIs into wherever the transaction is happening, be it within the AI assistant, embedded experiences like Social, or agentic commerce embedded payments.

I’d avoid locking yourself into one agent, as commerce is a fast-moving landscape in which you need the capability to interact across several touch points. So, independence via open, flexible infrastructure is paramount.

Then, consider the variables that determine how an agent evaluates your systems. Speed of response is important, but just one of many factors. Think about how agents “see” your offering. A human shopper views traditional merchandising, such as category pages, banners, and offers. Conversely, AI agents parse structured data, so you’ll need to provide meaningful information in natural language, explaining what the product is and how it works, to drive engagement.

You’ll also need to move from keyword optimization for search engine crawlers to providing a clear information architecture with imagery, descriptive headings, bullet points, and tables that agents can digest.

Meanwhile, agents will be looking for architecture that provides real-time inventory and transparent pricing data. Details of delivery and return processes need to be detailed, accompanied by clear customer support information.

Finally, a frictionless final payment system that can transact seamlessly with agents is fundamental. Smart, integrated payment structures bring benefits in efficiency, personalization, and scalability. For instance, PayPal’s agentic commerce toolkit helps businesses and developers create apps to manage complex payment workflows, which integrates with our own Shopware platform.

These factors all create signals that encourage agents to trust your systems and recommend your products. Strong long-term credibility with commerce agents will help ensure your brand surfaces regularly and you receive repeat orders.

This all entails a huge mindset shift across the industry; a move from thinking about conversion optimization to agentic decision optimization. But there are also huge opportunities for those who understand and act on how shopper behaviors are now rapidly changing. The question is, are you ready to change, too?

Jason Nyhus is the president and general manager of Shopware US, and an experienced sales and marketing leader with a history of success spanning over 20 years, driving significant levels of revenue growth and market adoption for leading technology organizations.