Topic starter
10/07/2026 9:09 pm
TL;DR: Retailers are moving conversational commerce into AI-driven discovery, checkout, and support, but external platforms still lack access to product, inventory, and customer context, according to Riskified. The identity question is no longer whether AI can sell, but whether merchants can keep decision authority, customer trust, and abuse controls inside their own environment.
Editorial analysis by NHI Mgmt Group, based on content published by Riskified: “Agentic commerce: harness it, don’t outsource it”.
Key questions
Q: How should security teams govern customer-facing AI without blocking useful interactions?
A: Put governance in the request and response path so the system can inspect prompts, classify intent, and apply policy before anything reaches the customer.
Q: Why do external AI platforms struggle with merchant-side decisions?
A: They usually lack the proprietary context that makes the decision safe, including product data, inventory, customer history, and behavioural signals.
Q: What are the failure modes when AI handles refunds and returns without identity context?
A: The system becomes either too generous or too restrictive.
Practitioner guidance
- Separate discovery from decisioning Use external AI as a referral or discovery layer only when it does not need to own the final customer decision.
- Feed identity intelligence into live workflows Integrate behavioural and historical identity signals into customer service and checkout automation at the moment of request so the system can distinguish trusted customers from serial abusers before action is taken.
- Define decision thresholds for high-risk requests Set explicit rules for when an AI agent can approve, defer, or escalate customer exceptions such as refunds, chargebacks, and missing-order claims, especially where financial exposure is immediate.
Bottom line: Conversational commerce is changing how merchants think about identity because the critical issue is no longer only who is asking, but who is authorised to decide.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
10/07/2026 9:32 pm
Merchant-side AI creates an identity control boundary, not just an experience choice. The article is really about where decision authority belongs when AI starts handling customer journeys. If the merchant does not keep the context, the AI cannot distinguish legitimate customers from policy abusers. Practitioners should treat conversational commerce as a governance boundary around who can decide, with what data, and under whose controls.
A question worth separating out:
A: Use external AI for discovery and referral when it adds reach, but keep the authoritative customer relationship and final service decisions inside the merchant environment. That split preserves brand control, protects first-party data, and lets the merchant apply its own abuse signals before money or service is committed.
👉 Read our full editorial: Conversational commerce shifts identity control to merchant-side AI