Infor connects Claude to manufacturing workflows

Infor connects Claude to manufacturing workflows

Infor and Anthropic have connected manufacturing systems to conversational AI. Selected customers can use the integration to investigate production problems and carry out authorised changes through existing business applications, with user approval required before records are modified.


Infor has introduced a manufacturing-focused integration with Anthropic’s Claude AI assistant that allows selected customers to investigate production and supply problems and then make approved changes inside Infor business systems. The 12 November partnership brings manufacturing workflows into conversational software without bypassing the permissions and process controls already configured in the underlying enterprise applications. Its initial availability is restricted to selected Infor Velocity Suite customers rather than the company’s entire installed base.

The product, Infor Industry AI for Claude, connects users to information about production orders, material supply, capacity and customer commitments. Production planners, buyers and customer service employees can request explanations in ordinary language, examine recommendations and authorise a follow-up action from the conversation. Infor provides the business-system connection and the operational rules; Anthropic supplies the Claude assistant through which the user can ask questions and review the results. The same functionality can also be accessed from supported Infor workspaces.

When a production order is in danger of missing its delivery date, a planner can ask the assistant to examine the constraint using the information available in Infor. A delayed material receipt, limited available capacity or a competing customer commitment may each lead to a different response. Purchasing staff can examine alternative supply and seek approval for a purchase order, while planners can review changes to scheduling or stock allocation. A proposed adjustment is carried out in the underlying Infor application rather than written to a separate record maintained by the AI assistant.

The connector checks the identity and operating permissions of the person submitting the request. Infor says single sign-on and role-based entitlements govern both the operational information returned and the actions that can be taken. Employees with different responsibilities may therefore receive different answers about the same order: a purchasing employee may be authorised to create or release a purchase order, while a production planner may have permission to change a schedule but not the supplier contract. The software does not grant new operational authority simply because a user can describe a transaction in natural language.

Infor’s stated safeguard is explicit approval before each change made through the Claude connector. Those actions remain subject to the company’s existing authorisation controls and are recorded in Infor. Stock reallocation illustrates the practical boundary: moving inventory to fulfil one urgent order may delay another, so the recommendation must be evaluated against current demand, available material and the employee’s authority. Infor’s published examples cover purchases, order recovery, inventory allocation and shipping activity; they do not establish autonomous control of factory machinery.

Infor is also providing manufacturing-specific vocabulary and logic. The integration is designed to recognise that an item, a stock-keeping unit and a product may refer to related entities in an operating workflow, while preserving the relationships held in the customer’s own systems. Questions about a product’s availability can involve stock at multiple sites, committed inventory, work in progress and expected receipts. A conversational answer that draws on only one of those values could be misleading, which is why the quality and coverage of the connected records remain central to the result.

The connector is built on the Model Context Protocol, an open standard intended to let software tools expose defined capabilities and information to AI applications. In this implementation, the protocol provides a route between Claude and authorised Infor functions rather than unrestricted access to the application’s database. The architecture may make it easier for users to work across compatible tools, but an organisation still has to configure the connector, control identity and check that each supported transaction corresponds to the right internal procedure.

Planning, purchasing, warehousing and production often maintain different portions of a manufacturer’s operating picture. A machine stoppage may reduce available capacity while a customer order still appears feasible in a separate business workflow. Infor’s connector is designed to surface the related business records and approved responses without requiring a user to piece them together manually across applications. It does not replace programmable controllers, safety functions or manufacturing equipment controls. The manufacturer’s own data quality and the consistency of the configured workflows remain limiting factors for any answer produced by the assistant.

The partnership sits alongside Anthropic’s broader introduction of Claude for Manufacturing, which includes skills and connectors for manufacturing-related tasks. That wider offering and Infor Industry AI for Claude should not be treated as identical products or as proof that every advertised workflow is already running at a customer plant. Infor has described the currently supported roles and examples, while additional skills, applications and roles are planned for subsequent releases. The extent of deployment and measured productivity outcomes have not been published in the joint announcement.

Customers using the initial service require their own Claude licence and an administrator must activate the Infor connector. The limited initial release creates a clear distinction between technical availability to an approved group and general rollout across manufacturing sites. Implementation will also depend on the customer’s existing system configuration, the quality of supplier and production data, and the decisions delegated to each employee. A useful answer to a production question cannot overcome inaccurate inventory balances or an unrecorded shop-floor disruption.

The immediate deployment remains limited to selected Velocity Suite customers. Infor has not disclosed a customer count, independently measured productivity improvement or date for broader availability. Further implementation will test whether planners and buyers can consistently resolve exceptions using accurate records and appropriate approvals, and whether new roles and skills can be added without weakening the existing division of operational responsibilities.


Stories for you


  • TIER IV develops driverless bulk transport at Japanese port

    TIER IV develops driverless bulk transport at Japanese port

    TIER IV will develop autonomous bulk handling transport in Japan. Its collaboration with Shunan Bulk Terminal aims to automate a defined industrial port route supplying Tokuyama’s power facilities, with driverless operation targeted for late 2028.


  • ASUS tests Gallop robot on electronics packing line

    ASUS tests Gallop robot on electronics packing line

    ASUS is testing Gallop on real electronics factory packing work. The fixed-base robot is learning to handle motherboard accessories, although its 99.5% success objective remains a target and customer pilots are not planned until 2027.