CustomerThink: Where is the Customer in Agentic Commerce?

Apr 6

This article was written by Greg Kihlström for CustomerThink. Read the full article here.

When purchases are made on behalf of customers without them ever seeing your carefully crafted display ad or landing page before the sale, what is the role of your brand in the transaction?

This is an immediate question to consider as we are witnessing the shift from a search-based economy to a machine-mediated one, where AI agents act as concierges that guide, or even entirely represent, shoppers.

The stakes for enterprise brands are high, with EMARKETER projecting that AI-driven e-commerce sales will surpass $144 billion by 2029. As platforms and marketplaces experiment with agentic models, the future shape of commerce is still forming. However, Radial’s recent research indicates that while consumers are curious, they are also hesitant.

Consumers are open to AI when it adds value, but are terrified of losing control over their data and actions. Marketing leaders who fail to adapt their strategy to this new agentic layer will find themselves invisible to the very tools their customers are starting to rely on.

Shifting From Emotional Connections to Data Credibility

For decades, the CMO’s playbook has been focused on creating emotional engagement and brand affinity. With the dawn of agentic commerce, that strategy must evolve to prioritize data authority and operational consistency. Since an AI agent cannot feel a connection to your brand, it must have other ways to assess whether you are the most logical, transparent, and reliable choice based on the parameters it has been given by the human relying on it.

The importance of trust does not go away, but its definition is shifting. Radial’s research shows that 62% of consumers will lose trust in a brand if product quality declines. In an agentic world, this expectation of quality extends to the digital experience itself. If your AI-facing data is messy, the machine deems your brand low-quality. Strategy must now focus on creating a positive feedback loop by listening to current customer anxieties to shape future agentic interactions. For this reason, dependability is key. Shauna Bowen, Chief Digital & Transformation Officer at Radial, adds, “agentic commerce shows great promise as AI-driven sales surge, but its prolonged success is dependent on reliability, not autonomy.”

Strategic success now requires balancing automation with the human need for oversight. Even when offered cost savings, Radial found that 34% of consumers want AI to take only actions they explicitly approve, according to Radial’s research. Your strategy cannot simply be to achieve complete end-to-end automation for many reasons. Instead, the focus should be on an assisted autonomy that respects the user’s desire for a kill switch.

Shifting to Optimize for Agentic Customers

When moving from strategy to tactics, marketing functions take on a more engineering-like role than ever. Previous tactics such as SEO, social media blitzes, and influencer partnerships are not enough when the shopper is an agent.

Instead, the following approaches will need to be adopted in the enterprise:

Your teams must ensure product data is digestible for emerging protocols like the Model Context Protocol (MCP) or Universal Commerce Protocol (UCP).
Since 64% of consumers worry about sharing payment info with AI, providing secure, known payment options is a marketing necessity.
Because 42% of shoppers expect to talk to a human if an AI fumbles, your approach must include a seamless escalation path to a person.
Tactics will likely need to vary by age and other characteristics of a segment. Boomers need human-to-human options (59%), while Gen Z focuses on brand transparency (27%).

These tactics represent a fundamental shift in team structure. You need data scientists and protocol experts sitting next to your copywriters. If the copy is great, but if your Order Management System can’t talk to the customer’s agent, the transaction dies. Marketing is now responsible for both the infrastructure and the positioning of the product.

Measuring Autonomous Success

If the customer is a machine, your traditional KPIs are likely misrepresenting the truth. A high click-through rate is meaningless if it comes from a bot that can’t complete the purchase due to a data mismatch. Success measurement must become more clinical and infrastructure-focused. Look at the following measures of success:

How often is your brand the top recommendation provided by major AI assistants?
What percentage of AI-initiated orders are completed without needing a human to fix a mistake?
How many milliseconds does it take for your inventory and pricing changes to reflect in the AI ecosystem?
Measuring brand health through the lens of security settings and privacy-first interactions rather than social sentiment.

We are moving toward a world where the conversion rate is a measure of how well your systems interact, rather than how well your creative influenced human behavior.

Identifying the Post-Purchase Opportunity

In the coming months, CMOs should look beyond the initial purchase. The real opportunity for AI currently lies in the post-purchase experience. 53% of consumers are likely to use AI for monitored delivery and problem-solving. AI assistants are better suited as “order concierges” today than as impulsive shoppers.

However, there is a catch. As AI companies begin implementing “agent fees” for in-platform purchases, transaction costs will rise. Retailers will face a choice: absorb these costs or pass them to customers who are primarily using AI to find deals. If the fee kills the “deal,” it kills the adoption.

Leaders must also monitor the competitive landscape of potentially overlapping protocols. Whether it is OpenAI’s AOC, Gemini’s UCP, or Anthropic’s MCP, your brand needs to be flexible enough to integrate with a variety of providers rather than betting on just one.

Summary

To win at agentic commerce, brands need to balance maintaining a human connection with their customers while mastering the art of talking to their machines. This requires a focus on technical flexibility, radical transparency, operational excellence, and unwavering data security. In the end, the winning approach may involve very much or very little automation. Bowen adds, “brands that are looking to see AI agent success and ROI need to focus on the basics of speed, transparency, and reliability, whether AI or humans power it.”

For brands, the challenge is balancing the very human needs of customers with the technical and data challenges that agentic commerce.