commercetools and Mirion Technologies co-design AI agent to cut B2B order intake time

commercetools announced on June 16 a co-design partnership with Mirion Technologies to develop an AI-powered B2B Intake Agent capable of converting unstructured customer order requests — emails, PDFs, spreadsheets, CSV files — into structured quotes and carts in minutes rather than hours or days.

The announcement, made in Boston, positions the capability as a direct answer to one of the more persistent friction points in enterprise sales operations: the manual re-entry of order data across multiple backend systems. According to DocuSign data cited by commercetools, sales representatives currently spend 70% of their working time on administrative tasks, a burden that slows response times, inflates costs, and creates friction for buyers.

How the intake agent works

Once an incoming request arrives — in whatever format the buyer sends — the agent extracts SKUs, quantities, and product specifications and matches them against catalog data. It then routes the request to the appropriate business unit, customer account, and pricing structure before generating a draft quote or cart ready for a sales rep to review.

The agent connects to CRM and customer service platforms such as Zendesk through an API-first architecture, fitting into existing enterprise workflows rather than requiring teams to adopt entirely new tooling. commercetools describes the underlying platform philosophy as "autonomous commerce," where AI systems execute operational decisions within defined business rules.

Many B2B businesses have talented sales and service teams spending too much time translating incoming order requests instead of serving customers. We built this capability to apply AI where it creates immediate value inside real workflows. It helps teams move faster today, while giving businesses a practical foundation for more agentic commerce over time. — Shiri Mosenzon-Erez, Chief Product Officer, commercetools

Why Mirion Technologies

Mirion Technologies manufactures radiation detection and measurement equipment — products that carry complex specifications, regulatory requirements, and procurement workflows where errors carry real consequences. That environment made the company a demanding proving ground for an AI order intake tool.

By co-designing the agent alongside Mirion, commercetools says it is stress-testing the capability against operational complexity that a simplified pilot environment would not surface. The stated goal is to ship something shaped by actual enterprise edge cases, not a theoretical average.

For manufacturers like Mirion, speed matters, but accuracy matters just as much. Many customer requests involve complex products, specific requirements, and workflows that cannot afford delays or errors. We see strong potential for AI to simplify order intake, improve responsiveness, and help our teams focus on solving customer needs. — Matthew Maddox, VP of Digital Commerce, Mirion

Broader implications for B2B sales operations

The case for automating order intake in B2B settings is not primarily about headcount reduction. Faster quote generation affects conversion rates, average order value, and customer retention — particularly in sectors where buyers evaluate vendors partly on response time and where orders are technically complex enough that a slow turnaround signals operational risk.

commercetools frames the Intake Agent as an entry point into a wider "autonomous commerce" strategy, where AI increasingly handles routine transactional decisions end-to-end. The Mirion partnership gives that strategy a reference customer with demanding real-world requirements, which enterprise buyers evaluating similar tooling will likely scrutinize closely.

The agent is broadly applicable across industries where B2B orders still arrive through offline or semi-structured channels — manufacturing, distribution, industrial supply, and healthcare procurement among them. For sales and operations leaders in those sectors, the central question will be how the system handles ambiguous or incomplete incoming requests, and what human-review steps remain before a quote reaches the customer.