Rethinking Conversational Commerce 

For years, we at Opus Research have explored the concept of conversational commerce—the idea that consumers could use natural language, through voice or text, to find products, place orders, and manage their shopping experiences. These systems recognized intent and returned results. They were reactive, responding to prompts like: 

“Keto-friendly meals for two.” 
“Order more paper towels.” 
“Where’s my package?” 

Enter Agentic Shopping 

But what we’re seeing now, especially with Walmart’s latest moves, is a gradual transition into something fundamentally different: Agentic Shopping. 

In her recent Forbes article, Kiri Masters describes Walmart’s vision for a retail world where AI agents act on behalf of users, not just responding to queries but executing complex goals across systems. Walmart’s new customer-facing agent, Sparky, is already live in their app and can help customers with their shopping.  

In her article, Masters quotes Hari Vasudev, CTO of Walmart U.S., as saying he expects the traditional search bar to be replaced by multimodal interfaces like Sparky. Instead of expecting the customer to search for specific items, these new intelligent agents invite customers to state their goals and have the agent do the leg work to determine what needs to end up in their shopping cart. 

The current version of Walmart’s Sparky isn’t quite up to the task yet, but more capable versions are in the works. Right now, if you tell Sparky: 

“My cousins are visiting next week and they’re both on a keto diet. Plan out 2 days of meals we can all enjoy” 

Sparky understands the goal, and with prodding can even plan out the meals. But for some reason it stops short of actually compiling the shopping list required to make the meals. Chances are good Sparky will get better at goal-driving planning in coming releases. 

The hallmark of agent-powered retail experiences, aka agentic shopping, is that these systems don’t just interpret requests. Instead, they maintain context over time, access backend inventory and logistics systems, automate multi-step tasks, and learn and adapt to user preferences. 

We are now seeing retail systems evolve from reactive Q&A bots to proactive, task-oriented agents that understand nuanced human intent. 

Preparing for BYO-Agent Commerce 

As Masters points out, Walmart is preparing not just to build its own agents, but also to operate in a future where consumers might bring their own agents to do the shopping. 

These third-party agents, whether built by OpenAI or others, will understand their users deeply and act autonomously on their behalf. They may interact directly with Walmart’s systems, comparing prices, applying loyalty benefits, or optimizing delivery preferences. 

This emerging model is what Walmart refers to as agent-to-agent commerce—a paradigm shift where the “customer” is no longer a person, but software acting on behalf of a person. 

Of course, a world without login pages will need rock-solid bot authentication. As consumers delegate more decisions to autonomous agents, retailers must ensure those agents are trustworthy, verifiably linked to real users, and granted only the permissions they need, preserving both privacy and control in the process.

Implications for Retailers and Brands 

This shift to agent-to-agent commerce, even though it’s not quite here yet, raises significant strategic questions for retailers. For one: how do retailers earn loyalty in a world where personal agents mediate the relationship? From a more operational standpoint, how should product pages, pricing structures, and promotions evolve to be machine-readable and agent-optimized? 

We should be accustomed to the pace of changes these days, but it’s still tough to image what customer experience could look like in a world with no search bars. 

From Dialogue to Delegation 

At Opus Research, we once defined conversational commerce as the use of natural language in digital shopping. Today, that idea has matured into agentic commerce, where consumers don’t just talk to retailers, but potentially even delegate to software. 

Walmart’s aggressive investment in agent infrastructure is a clear signal that they understand the shift. They’re going beyond simply enabling conversation to preparing for negotiation, delegation, and autonomy. 

Retailers who don’t begin to prepare for this transformation risk being invisible not only to their customers, but to their customers’ agents as well. 

Categories: Articles