Anthropic schedules November 17 webinar on Claude’s role in factory work

Anthropic will hold its Claude for Manufacturing webinar on November 17, 2026, from 8:00 to 9:00 AM PT. The official event page offers a registration form, but no direct watch link. The one-hour technical session will cover drafting plant paperwork and a research preview for connecting AI agents to physical equipment.

Registering for the November session

The event is titled “Claude for Manufacturing: Put Your Plant’s Know-How to Work.” Anthropic lists the full hour as the technical webinar, rather than identifying a separate keynote or doors-opening time. Its page invites attendees to reserve a place through the form. A direct stream address is not supplied, so that registration page is the official starting point—not a verified public broadcast link.

The plant work Anthropic will demonstrate

The confirmed agenda combines work Anthropic says teams already perform with Claude and a look at equipment integration research. The workflow portion centers on records plants already keep, rather than asking attendees to begin with a new collection of information. Anthropic says it will show how teams adapt Claude’s drafts to their own forms and procedures.

• Existing systems: Teams and partners will discuss work in their own enterprise resource planning, computer-aided design and document systems.
• Equipment research: The session will cover the Model Hardware Standard, its research-preview status, how equipment communicates capabilities and limits, and how to apply.

The approval boundary is explicit. Anthropic says quality, estimating and plant teams approve the documents before they go anywhere. Nothing writes independently to production control or a system of record—the official system holding operational information. That describes a drafting-and-review workflow, not permission for the paperwork demonstrations to make autonomous changes to plant operations.

How the hardware interface works—and where it stops

In its earlier research-preview announcement, Anthropic described the Model Hardware Standard, or MHS, as a shared specification for AI agents to operate devices with programmable interfaces. It is model-agnostic, meaning it is not restricted to Claude. The aim is to replace separate, custom connections between equipment and agents with a common way to discover and operate devices.

The mechanism is a standardized driver: software that translates instructions for a hardware device. It uses commands such as read and write, while natural-language tags capture characteristics that code alone may not reveal. Anthropic gives a robot arm’s weight as one example. The driver produces a reference file describing what a device can measure, what can be adjusted and which safety limits will be enforced.

Those descriptions do not remove the need for physical expertise. Anthropic says Claude’s spatial and physical reasoning still requires expert oversight. In one example, Genentech researchers had to help Claude recognize that foaming in protein samples was a physical failure, not a software bug. MHS also does not yet work with equipment that lacks a programming interface.

For attendees evaluating the paperwork demonstrations, Anthropic promises a concrete assessment route: judge the work on a process the plant already measures.