Benjamin Fabre is the cofounder and CEO of DataDome.
Right now, across boardrooms and strategy meetings, e-commerce leaders are debating whether they should allow AI agents to access their sites.
But here’s the problem: That debate is already outdated.
AI agents aren’t coming to e-commerce. They’re already here. And most businesses have no idea what’s happening on their own digital properties.
According to recent analysis by my company, DataDome, AI agents now account for 10% of verified bot traffic, with 64% of those agents hitting form pages—clear signals of commercial intent. Yet many e-commerce leaders remain unaware of this shift, largely because their analytics tools weren’t built to distinguish AI agents from generic bot traffic.
While companies argue over access policies, AI agents are already navigating their product catalogs, evaluating their checkout flows, and making recommendations that shape billions in purchasing decisions.
The real question isn’t whether to allow AI agents; it’s whether you can afford to remain invisible to them.
The Commerce Shift Is Already Underway
The numbers tell a story that most executives haven’t yet processed. Ark Investment Management projects that agentic AI will drive $9 trillion in e-commerce transactions by 2030, representing roughly 25% of global online commerce.
Major platforms have already deployed agent-based commerce features: Google is piloting agentic checkout systems, Amazon launched “Buy for Me” functionality, and Perplexity offers free agent-assisted shopping experiences.
However, traditional tools lump AI agents into catch-all categories like “unknown bots” or fail to identify them entirely. The result is that business leaders are making strategic decisions about one of the most significant shifts in digital commerce without any visibility into what’s actually happening on their sites.
What You Can’t See Will Cost You
This visibility gap creates concrete business risks that extend well beyond missed opportunities.
If your analytics can’t tell you which AI agents are visiting, which product pages they’re examining or why they’re accessing your forms, you’re essentially operating blind. You can’t optimize for what you can’t measure. You can’t protect against threats you can’t identify. And you certainly can’t capture opportunities you don’t know exist.
The competitive dynamics are equally stark. While you debate access policies, your competitors may be structuring their product data for machine readability, exposing APIs for agent integration, and optimizing their sites for AI-mediated transactions. When an AI agent evaluates options on behalf of a consumer, inclusion in that evaluation isn’t guaranteed. Rather, it’s earned through discoverability and accessibility.
If your product catalog isn’t machine-readable, if your pricing isn’t exposed through structured markup, if your site architecture doesn’t support agent navigation, you simply won’t appear in the recommendation. The consumer never sees your brand. You’ve lost the sale before you knew there was an opportunity.
There’s also a security dimension that can’t be ignored. AI agents introduce new attack vectors, from credential stuffing at machine scale to agent impersonation and logic manipulation. Without the ability to distinguish legitimate agent activity from malicious traffic, security teams are fighting yesterday’s battles with yesterday’s tools.
From Debate To Decision
The path forward requires a fundamental reframing. The question isn’t whether to allow AI agents, it’s how to enable the right ones while protecting against the wrong ones.
That shift begins with visibility. Before you can formulate an agent strategy, you need to understand your current agent activity: Which agents are accessing your site? What pages are they viewing? What actions are they attempting? What patterns distinguish legitimate commercial agents from scrapers or fraudsters?
This isn’t a theoretical exercise. The data exists in your traffic logs right now. But without tools designed to identify and classify AI agent activity, it remains invisible, buried in aggregated metrics that mask the signals you need to make informed decisions.
Once you have visibility, the strategic choices become clearer. You can identify high-value agent traffic that represents a genuine commercial opportunity, optimize your site architecture and data structures to support legitimate agent interactions, implement access policies that balance openness with protection, and detect and block malicious agent activity before it impacts your operations or your customers.
Questions Every E-Commerce Leader Should Be Asking
If you’re responsible for digital commerce strategy, it’s time to pressure-test your organization’s readiness. Start by asking whether your team can answer these fundamental questions:
• Can you identify which AI agents are currently accessing your site and distinguish between commercial agents, research crawlers and potential threats?
• Do you know which product pages or transaction flows agents are examining most frequently?
Beyond visibility, consider your competitive positioning:
• Is your product data structured in machine-readable formats that agents can easily parse and evaluate?
• Are your pricing, availability and specifications exposed through APIs or structured markup that agents can access programmatically?
Finally, examine your security posture:
• Can your current fraud prevention tools detect anomalous agent behavior in real time?
• Do you have the capability to block malicious agent traffic without disrupting legitimate agent-mediated commerce?
• And perhaps most critically, do you know what you don’t know about the agent activity happening on your digital properties right now?
If you can’t answer these questions with confidence, you’re not alone. But that doesn’t make the gap any less urgent to address.
The Window Is Narrow
A decade ago, businesses that failed to invest in search engine optimization found themselves invisible to consumers. The same dynamic is now playing out with AI agents, but the timeline is compressed.
The companies that establish visibility into agent activity today will shape how those agents interact with their businesses tomorrow. They’ll optimize for discoverability, build integration points for high-value agents, and create defensive measures against malicious traffic. They’ll capture market share in agent-mediated transactions while competitors remain locked out of recommendation engines they don’t even know exist.
The companies that wait will find themselves addressing a problem that’s already cost them revenue, customer relationships and competitive position.
The era of agentic commerce has begun. The only question is whether you’ll participate in it with intention and strategy, or discover too late that the market moved while you were still debating whether it would.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
