Critical Manufacturing announced Version 12.0 of its MES-powered Industrial Operations Platform on October 20, 2026, adding a catalog of purpose-built AI agents, a module for keeping master data consistent across plants, and role-based interface views. CEO Francisco Almada Lobo framed the release around context rather than raw data volume: "Manufacturers don't need more data, they need intelligence that understands their operations and can support decisions" within them. The platform targets semiconductor, electronics, medical device and industrial equipment manufacturers, according to the company's press release and coverage by Electronics Media.
Two Kinds of Agent: Ask and Build
Gen AI Ask Agents let operators and supervisors pose questions in plain language — about KPI trends, shift handovers, or why a line's output dropped — and get answers drawn from the platform's Canonical Data Model, so the response reflects actual product definitions, resources and process flows rather than a generic chatbot guess. Gen AI Build Agents go a step further: they generate dashboards and interface pages directly from a description of what a user needs to see, without routing the request through a developer backlog.
Critical Manufacturing Version 12.0 — What Changed
- Gen AI Ask Agents — natural-language questions answered from the platform's governed production data, built for scenarios like KPI analysis and shift handovers;
- Gen AI Build Agents — generate dashboards and UI pages as governed, versionable configuration rather than generated code, refinable by conversation or manual edit;
- Global Data Orchestrator — add-on module for master-data consistency across multiple factory sites, with change packages and dry-run validation before rollout;
- User Workspaces — role-based views for operators, supervisors and quality managers meant to cut navigation complexity and training time.
Why "Governed Configuration" Is the Detail That Matters
The distinction Critical Manufacturing draws — output as governed configuration, not generated code — is the part worth paying attention to. A dashboard assembled by an AI agent still has to be versioned, validated and auditable the same way a hand-built screen would be in a regulated plant; a shop floor running a Part 11 or ISO 13485 process can't accept a UI element nobody can explain six months later. That is exactly the gap between a quick AI demo and an interface a quality manager will actually sign off on, and it is the same gap a team that designs and validates operator-facing interfaces for regulated manufacturing environments is built to close — turning an agent-generated layout into something consistent with the rest of a plant's workspaces, legible under shift lighting, and traceable for an audit.
Role-Based Workspaces Reflect a Wider Trend
User Workspaces — distinct views for a floor operator versus a quality manager versus a supervisor — echo what other MES and HMI vendors have been shipping through 2026 as AI reshapes what an operator screen needs to surface: prioritized alarms, maintenance recommendations and context for why a process changed, rather than a flat table of tags. Pairing that interface layer with a Global Data Orchestrator that validates master-data changes before they roll out to every site addresses a separate, older problem: a multisite manufacturer whose plants quietly drift apart on product definitions and routing data until a dashboard shows numbers that don't reconcile.
Sources: Critical Manufacturing press release, October 20, 2026; Electronics Media, October 27, 2026.