The agentic commerce radar: Spring 2026 update

The state of agentic commerce

Agentic commerce continues to make waves (not to say a tsunami) across the retail industry, with the latest developments certainly marking a shift from hype to reality. In such a short period of time, we started to see all things agentic and AI in action, from Google pushing open protocols such as UCP to brands like Sephora treading their own AI path. 

Meanwhile, consumers are becoming more accustomed to navigating the AI landscape like pros, increasingly leveraging their favorite AI channels to discover and compare products, but remain wary of giving agents autonomy to buy on their behalf. Indeed, autonomous shopping hasn’t even started yet, and it may take even longer than anticipated to make this a reality, considering that OpenAI recently paused its Instant Checkout. 

In the short term, this translates into a “simple” requirement for brands and retailers: Make product catalogs, inventory, payments and trust signals readable by machines as well as humans. But in the long term — when machine-readable data becomes the entry-level to compete in the agentic marketplace — the winners will be those with the foundation to leverage agentic across AI channels and within their brand-owned touchpoints. 

Let’s explore in more detail what’s happened in agentic commerce over the last couple of months — and what this means for your business.

Consumer sentiment for all things AI continues to evolve

What happened? 

• 63% of European shoppers use AI for comparing options like brands, models, prices and reviews, and 55% use AI to learn about a category or product. At the same time, 53% are comfortable letting AI manage recurring purchases automatically. 
• 51% of shoppers would be open to AI handling the entire shopping process, including making the final purchase.
• 62% of Australian consumers are open to AI agents helping them make purchasing decisions, particularly for everyday categories such as household items, subscriptions, gifts and even travel bookings.
• About 65% of US consumers trust AI to compare prices, but only 14% trust it to place orders on their behalf. Trust is highest among Gen Z (29%) and millennials (30%). 
• 79% of consumers say accuracy is the most important factor in AI-powered shopping. 

Why this matters

Consumer adoption of AI in commerce is no longer the question; it’s already happening at scale. At the top, AI is quickly becoming the default interface for research, discovery and comparison, but the adoption is uneven across the funnel. 

The gap between 65% trusting AI for comparison vs. 14% for purchasing is the most important signal in this data. It shows that consumers trust AI to inform decisions at this moment, and not yet to make them. While this behavior resonates across all demographics, younger consumers are more willing to delegate decisions. This suggests that the shift to agent-led commerce will be gradual, driven by repeated positive experiences.

To ensure customers consistently have positive experiences, data accuracy is critical — and consumers know it. Before AI can transact, it must consistently demonstrate that its recommendations are reliable, unbiased and verifiable. 

AI-assisted shopping is here. Transaction execution, not so much

What happened?  

• OpenAI paused Instant Checkout in ChatGPT, shifting focus from in-chat transactions to product discovery and comparison.
• Walmart saw 3x lower conversion rates for in-chat purchases vs. redirecting users to its own website.
• Google launched the Universal Commerce Protocol (UCP) to enable agent-to-agent transactions across retailers, payments and AI systems.
• Amazon expanded its “Buy for me” capability and evolved Amazon Rufus toward comparison and autonomous purchasing.
• Amazon and OpenAI announced a major strategic partnership, signaling deeper alignment between AI and commerce infrastructure.

Why this matters

As mentioned in the previous section, AI-assisted shopping has clearly crossed the adoption threshold. Consumers are increasingly comfortable using AI to search and compare products. While the top of the funnel continues to shift, it’s the execution of transactions where reality hits.

The pause of Instant Checkout is the clearest signal, as the most well-known AI platform is refocusing on experience quality, reliability and enterprise readiness. In other words, the ambition didn’t change, but the infrastructure and trust layers aren’t ready yet. At the same time, the deal with Amazon and OpenAI may change the course for agentic checkout, but it’s too early to say what will happen and how. 

Walmart’s conversion data reinforces that ChatGPT’s Instant Checkout wasn’t ready. A 3x drop in performance is a behavioral signal that shows consumers are still more comfortable completing purchases in known environments, even if discovery starts elsewhere. 

But there’s more in agentic commerce than OpenAI’s ups and downs. Google’s UCP and Gemini are building the connective tissue between agents, merchants and payment providers. This shifts control back toward merchants and ecosystems rather than a single AI interface. 

Amazon, meanwhile, is doubling down on its strength: Closed-loop commerce. With Rufus and “Buy for me,” the entire journey can be compressed because it already owns fulfillment, payments and trust.

The result is a split market: Some players are trying to own the transaction, others are trying to orchestrate it, and consumers are still deciding whether they’re ready to delegate decisions to an AI agent. 

Brand-owned ecosystems are claiming customer control and ownership

What happened?

• Gap Inc. partnered with Google’s Gemini to enable agentic shopping while retaining greater control over the customer experience. 
• After underperformance in ChatGPT checkout, Walmart integrated its own AI assistant, Sparky, into ChatGPT, with plans to expand to Gemini. 
• Walmart reports that ChatGPT is driving ~2x as many new customers as traditional search channels. 
• Sephora launched an AI-powered shopping experience within ChatGPT, starting in the US, focused on personalized recommendations and guided discovery.

Why this matters

The first wave of agentic commerce suggested that platforms like ChatGPT or Gemini could disintermediate brands entirely, owning everything from discovery to checkout. But early data have changed that narrative.

Walmart’s pivot is especially telling. Instead of relying on platform-native checkout, it embedded its own intelligence (Sparky) into the AI layer while keeping the transaction anchored in its own ecosystem. The basket, the customer data and the relationship remain under Walmart’s control, even if the interaction starts elsewhere.

Sephora is taking a similar approach, using AI within platform environments to enhance discovery and personalization while still tying the experience back to its app, loyalty programs and product universe.

Even Gap’s partnership with Gemini reflects this shift. Compared to more closed systems, some platforms are positioning themselves as infrastructure layers, giving brands more influence over how their products are presented and sold.

In this model, control doesn’t come from owning the interface; it comes from owning the data, logic and transaction layer behind it.

Deal maker or breaker? Security, trust and identity

What happened?

• Visa Intelligent Commerce initiative announced that hundreds of agent-initiated transactions had been successfully completed in real-world controlled environments with partners across its network. 
• Visa opened integration for AI agent payments. 
• Mastercard introduced Verifiable Intent, a new trust paradigm for agentic commerce, co-developed with Google. 
• AI-driven fraud is accelerating, including voice impersonation attacks, automated phishing at scale and synthetic identities controlled by AI agents.
• Only 24.4% of organizations have full visibility into AI agents, and more than half of the agents run without security oversight or logging. 
• The same survey notes that 88% of organizations had confirmed or suspected AI agent incidents in the past year.
• While Responsible AI maturity is rising, only about one‑third of organizations report maturity levels of 3 or higher in strategy, governance and agentic AI controls.

Why this matters 

Traditional commerce relies on verifying a user’s credentials, devices and authentication flows. But in an agent-driven world, those signals no longer cover the AI agent. The system must identify who the agent is, whether it’s authorized to perform a particular task, and whether it’s acting within its intended scope. 

This is an industry challenge that highlights a structural risk in the current wave of agent adoption and its commercial impact. 

The data shows a widening gap: AI adoption is accelerating, but security readiness is lagging behind. Big players like Okta, Visa and Mastercard are creating solutions that tackle the problem, but cohesion is missing, including consistent identity models, centralized enforcement, clear ownership and continuous visibility. 

The foundational readiness for the agentic enterprise

Agentic commerce isn’t unfolding as a single, linear shift toward autonomous checkout. It’s emerging as a multi-layered ecosystem, where discovery, decision-making, transactions and trust are evolving at different speeds. 

In the short term, AI is already reshaping how customers find and evaluate products, making visibility, accuracy and accessibility in AI-driven environments the immediate priority. But as protocols mature and agents become more autonomous, competition will shift toward those who can orchestrate the entire experience across platforms and brand-owned ecosystems alike.

This is where foundational readiness becomes essential. Agentic commerce promises automation and personalization at scale, but it also exposes every weakness in your stack. AI doesn’t fix fragmented data, disconnected systems or inconsistent experiences; it amplifies them. Poor product data leads to flawed recommendations. Siloed infrastructure results in broken journeys. Disconnected checkout and loyalty flows turn automation into accelerated abandonment.

Foundational readiness is what turns AI from experimentation into execution. The brands that win will adopt agentic capabilities, powered by an infrastructure that supports them at scale without compromising performance or trust.