Building a Zero-Trust Gateway for Industrial AI

One problem I'm exploring:

How do you safely connect AI Agents to physical industrial systems?

Connecting an LLM/agent directly to SCADA, MES or robotics creates a very different risk profile from a normal API integration.

So I built the concept of SEOSiri Industrial AI Gateway (IAIG) around a governed intermediary layer.

The architecture brings together:

ISA-95 Unified Namespace
SCADA/MES protocol bridges
Digital Twin validation
AI Firewall & access control
Deterministic guardrails
TPM 2.0 command signing
ROS 2 robotics integration

The execution path becomes:

AI decision → Validate → Secure → Sign → Execute

I'm particularly interested in the architectural question:

What should the trust boundary look like when an AI agent can influence a cyber-physical system?

The full architecture and implementation details are available in the project write-up.

https://www.seosiri.com/2026/08/industrial-ai-gateway.html

Would love to hear how others are approaching AI + industrial automation + safety.