Agentic Commerce Hasn't Had Its ChatGPT Moment

Building trust in agentic commerce: Where brands should begin

The infrastructure for AI agents to browse and make purchases on behalf of humans is being built. Consumers are curious, some even enthusiastic, but the vast majority of us are not yet ready to hand over our wallets to an unproven technology. The smartest place for brands to start looks a lot more familiar.

How businesses can implement agentic commerce strategies to enhance customer engagement

When ChatGPT launched in late 2022, it didn’t just introduce a new capability, it finally gave consumers a mental model that they were comfortable with for how to interact with AI. The ease of the chat interface unlocked mass adoption in a way that years of prior AI work had not.

Agentic commerce (the idea of AI agents acting on your behalf to discover, evaluate and purchase products and services) has not yet had that moment. But the infrastructure is under construction. Visa and Mastercard are building agent payment rails while Amex is developing frameworks for autonomous transactions. But we are still waiting for the on ramp that placates consumer concerns and makes us feel comfortable giving an agent our credit card details.

Adobe’s 2026 Digital Trends Consumer Report, a global survey of 4,000 consumers, quantifies that gap.

22%

Early adopters
First to try new tools. Use AI regularly—more than double the survey average.
~25% already have their own AI agent.

61%

Secondary adopters
Open to new technology when it proves useful but prefer to wait for broader adoption.
Only 7% have their own AI agent.

17%

Holdouts
Avoid new technology whenever possible or only use it when they have no choice.
Majority oppose creating an AI agent.

That tells us that the consumers who are genuinely agent-ready today are a minority within a minority. Even among the curious, Adobe’s data reveals a significant gap between interest and intent.

43%

of consumers would be willing to try a brand’s AI personal concierge or agent if the service is offered.

19%

want a future where AI agents become their primary way of interacting with brands.

This gap between those willing to try AI agents and those using them as their default option is the trust and experience gap the industry needs to close. But closing this gap is not a technology problem; it’s a design problem.

What early agentic commerce experiments reveal about consumer behavior

OpenAI tested this boundary already and learned from it. In September 2025, the company launched Instant Checkout inside ChatGPT with Etsy and Shopify merchants, with Stripe powering the transaction layer. The premise made a lot of sense. Users already ask ChatGPT what to buy, so why route them somewhere else to pay? Six months later, the feature was quietly shelved. Users browsed in enormous numbers; they just didn’t buy. Etsy, one of the first merchants to participate, confirmed it saw no meaningful sales volume from the feature.

Forrester’s State of Agentic Commerce, 2026 reinforces why: among US consumers who regularly use answer engines, completing a purchase within the answer engine was the least-adopted use case. Coming in well behind asking general questions and researching products. 

“The discovery layer worked. The transaction layer did not. This isn’t a failure of agentic commerce—it’s a signal about sequencing.”

–Forrester (2026)

In no way does this signal that OpenAI is abandoning commerce. These insights have revamped ChatGPT’s shopping experience to focus on discovery and comparison, which is an acknowledgment that this is where consumers are actually ready to engage.

The brands and platforms that will win in the agentic commerce era are those that use this window to earn trust, not assume it.

What is agentic commerce and how does it differ from traditional e-commerce?

While the industry debates the future, a commerce shift is already delivering real results by making use of a modality that consumers already enjoy. Agentic commerce will go far beyond conversation, but conversational experiences are emerging as the natural first step. This is because they translate emerging capability into behavior consumers already understand and embrace.

Using Google’s Gemini Enterprise for CX, Macy’s built “Ask Macy’s,” their multimodal AI shopping concierge, in just four weeks. The agent handles natural language queries, offers a virtual try-on feature and responds to nuanced prompts, such as ‘I’m petite, I value natural fabrics and I need something that works from the office to a dinner engagement.’ Early beta data shows revenue per visit was roughly 4.75 times higher among customers who engaged with Ask Macy’s compared to those who didn’t. This is a significant improvement in what the digital retail experience can deliver.

In the home furnishing sphere, IKEA deployed a conversational AI agent, “Billy,” to handle routine customer queries. Billy now handles around 57% of interactions without human intervention. For the remaining 43%, IKEA learned that customers wanted advice–mostly related to interior design. So, Ikea reskilled the service workforce, turning employees into design consultants. The resulting paid consultancy offering has generated close to €1 billion in additional revenue.

Walmart has taken a different angle, embedding its “Sparky” assistant into its app. The company has also partnered with both OpenAI and Google to bring its product catalog into the discovery experience within ChatGPT and Gemini. In an internal survey, 81% of Walmart customers said they used Sparky to check product availability or review specifications before purchasing. Tellingly, this is a behavior that is anchored in research rather than transactions.

The discovery problem: don’t automate the best part

One of the most important tensions in agentic commerce is that some of what we’d term the friction in shopping isn’t friction at all; it’s the fun part, building anticipation and reinsurance.

For many of us, the joy of a vacation begins long before the plane leaves the runway. A two-week stay becomes a month-long experience simply by planning the places we want to visit weeks earlier. We actually want to compare all those cruises and theme parks to make sure we’re not missing out on something better. But if an agent could book them for us in the blink of an eye, are we eliminating the anticipation and the deliberation that is part of the joy?

A well-crafted conversational commerce experience that is a first step toward agentic commerce can optimize the process while retaining the anticipatory magic. Instead of collapsing discovery into a transaction, it can enrich it, weaving in imagery, video and dynamic UI that creates something no static page can match. The goal isn’t efficiency alone; it’s discovery that feels as compelling as the purchase. And crucially, Adobe’s data shows that the option to hand off to a human remains important across all consumer types, even for early adopters. That preference for a safety valve is a design brief, not an obstacle.

Imagining new discovery experiences, each distinctly shaped by what a brand stands for and uniquely offers, will be one of the most effective ways to stay close to consumers and avoid full agentic disintermediation.

From push to pull: redesigning for the agentic era

There’s a broader strategic shift embedded in all of this. For the past decade, personalization has been a fundamentally push model: brands tracked, listened and followed consumers, serving content that was often as creepy as it was useful. The era of AI-native marketing and agentic commerce inverts this. Customers call on the brand when they are ready. Today, that happens through a conversational interface; tomorrow, it will be through an agent. The brand’s job is to be findable, legible and compelling when that moment arrives. This requires brands to think differently about readiness: less of ‘how do we reach customers’ and more of ‘how do we make ourselves worth reaching.’ 

Conversational commerce is the first practical place to build that muscle and to build the trust infrastructure that agentic commerce optimization will eventually require.

What step one actually looks like

For brands ready to embrace agentic AI, the entry point is more accessible than it might seem. Replacing or augmenting on-site search with a conversational layer that handles discovery, answers nuanced questions and surfaces personalized content is buildable today. The design challenge, making it feel genuinely useful rather than like an overgrown chatbot, is the real challenge for brands. From there, a natural progression emerges: dynamic UI that responds to conversation context, integration with warm handoffs from LLM searches, contextual templates that adapt to different customer intents, and, ultimately, the payment and trust infrastructure that will underpin true agentic commerce protocols.

The brands that will be ready when agentic commerce finds its ChatGPT moment are the ones currently building the foundation. This is not because they are chasing a trend, but because conversational commerce is genuinely better for their customers today.

We think that Dynamic or Contextual UI has a big role to play in the future of both Agentic Commerce and AI Native Products and Services. Stay turned for next article, we will write more about this.