Retailers have spent years getting better at spotting bots. Now some of those bots may be customers, shopping partners, or the systems deciding whether to recommend products to customers. But you still need to block those that would spam or otherwise harm or risk your digital properties and data. In short, your job managing bots just got a lot harder, as:

  • Blocking too aggressively risks shutting off retail experiences from emerging AI agent-led customer journeys.
  • Allowing too freely invites familiar fraud and scraping risks through a much wider door.

Digital business and tech leaders, keep reading: Yes, bot and agent trust management historically sits with the CISO organization as a security conversation. But agentic commerce makes it a very real commerce conversation, too.

Understand “Who” AI Agents Really Are As They Become Part Of The Customer Journey

Our 2027 digital commerce predictions describe a future in which visibility and influence extend beyond retailers’ owned experiences. Consumers delegate decisions to agents, while service providers build infrastructure for agentic commerce. As such, bot and agent trust management therefore becomes one piece of the infrastructure connecting a retailer to that external ecosystem.

What about today? Relatively few consumers use AI shopping agents today, though Meta says shopping is the most common use case for Muse. Instinct’s founder says it is approaching over a billion dollars a year in transaction volume through the platform, despite the small customer base.

Additionally, 80% of online adults in the US, UK, and Canada who have used AI tools to shop say they use the tools to compare products they might like to buy (but not necessarily to complete that purchase). Retailers and brands must make their product information visible and available to these systems if they want to appear in those recommendations.

The upshot: retailers today already face automated visitors whom they actively want to serve, such as:

  • An agent scraping product data on behalf of a marketplace, such as Amazon gathering data for its Buy for Me portion of the catalog.
  • An answer engine like ChatGPT or Google Gemini accessing product content so that it can surface products in synthesized answers to customers.
  • A consumer’s own AI shopping agent (e.g., Instinct or Muse), acting as a digital delegate for the consumer in product discovery or purchase.

Beyond Just Blocking or Allowing, Transparency Provides Needed Nuance

Retail technology leaders need to manage automated traffic as a spectrum of actors and intentions rather than a single unwanted category. In good news, there are modern solutions to help you do exactly that.

In Forrester’s Buyer’s Guide: Bot And Agent Trust Management Software, 2026, we interviewed 22 reference customers. Of those, 17 used bot and agent trust management software for account fraud, 14 for e-commerce fraud, and 14 for web and LLM scraping. Thirteen had started or were about to start applying this software to AI agent trust.

Bot and agent trust management supports several jobs simultaneously: establishing transparent, trusted relationships with customers directly and through agents, detecting malicious automated application traffic, and identifying and reporting on the intent of AI-agent traffic.

Similarly, AI assistants in commerce software demand visibility, traceability, appropriate authority, and explainability. All of this creates transparency as well. Hidden decision-making becomes an operational liability quickly.

What Retail Technology Leaders Should Do Now

  • Stop treating every bot as the same problem. Define policies for malicious automation, known partners, answer engines, commerce agents, and customer-authorized shopping agents. Make explicit the business owner and desired behavior for each category. This foundational step builds on the emerging need that Forrester’s research has identified to understand AI-agent traffic and allow appropriate LLM-partner traffic.
  • Demand visibility and explainability from your vendor(s). Require bot and agent trust management vendors to provide the context behind classifications and decisions, not simply an allow/block result. Determine whether your teams can access enough signal and decisioning data to investigate incidents and understand why legitimate agents are allowed or denied.
  • Test for false positives where commerce is headed, not just where it has been. Include legitimate AI-agent scenarios in proofs of concept. We recommend putting prospective vendors through POCs and examining their detection strength, data, support, responsiveness, and ability to tune models as threats evolve.
  • Bring commerce and security leaders to the same table. Commerce teams need to decide which external agents they want to serve and what those agents should be able to do. Security teams need to determine how that access can be trusted and governed. As more shopping happens outside retailers’ traditional storefronts, these two decisions become inseparable.
  • Evaluate the vendor, not just its detection engine. Threats and agent ecosystems will change faster than any static implementation. Assess the vendor’s threat research, tuning capability, support, responsiveness, pricing model, and service options. Customers in Forrester’s bot and agent trust management bot and agent trust management research particularly valued strong vendor support and treated some relationships more like partnerships than conventional software contracts.

Get in touch with us (Emily on Commerce and Sandy on bot and agent trust management) to talk more about how Agentic Commerce brings these two disciplines together.

Share