AVEVA Maps a Safer Path to Autonomous Industrial AI
AVEVA maps safeguards for autonomous industrial AI as adoption rises nearly 78%, balancing plant safety, workforce change, grid resilience and sustainability.
Summary
On October 8, 2026, AVEVA chief technologist Arti Garg said industrial AI is shifting from more than 20 years of predictive analytics toward foundation models, physical AI and agentic AI, with one study indicating adoption rose almost 78% in two years. These systems can correlate equipment telemetry, service logs and engineering documents through graph databases and AI, giving operators near real-time diagnostics while robots collect data in hazardous areas. Because evolving, less explainable models can directly change mission-critical equipment, AVEVA prioritizes security, environmental efficiency, human safety and oversight. Humans remain supervisors in critical loops, with limits such as permitted operating bands; customer pilots have already progressed from recommended equipment set points to automatic adjustments.
Industrial AI could preserve knowledge as almost half the workforce is set to retire within five years, transfer experience between sites and help manage distributed renewable generation. AVEVA is working with Idaho National Laboratory’s AI grid resilience project on capacity, load, power quality and anomaly monitoring. Garg chairs IEEE’s P7100 working group, launched a little over two years ago, to standardize AI impact measurement across electricity, energy, resources, water and carbon; she also advocates right-sized, purpose-built models.
Thailand’s SCG Chemicals combined operational and engineering data with AVEVA Predictive Analytics’ proprietary anomaly models, targeting 99% plant reliability after early pilots delivered nearly 9x ROI by converting unplanned downtime into planned maintenance. Within 18 months, Garg expects domain experts to use AI-assisted coding to build applications. Wider robot and drone autonomy will require sandboxed testing, redesigned business processes and application-specific guardrails in plants, power systems and mines. The discussion was sponsored and produced in partnership with AVEVA.
Positives
- Almost 78% industrial AI adoption growth over two years signals rapid uptake of foundation, physical and agentic technologies.
- SCG Chemicals’ early pilots delivered nearly 9x ROI and support its target of 99% plant reliability.
- AVEVA and Idaho National Laboratory are testing AI to improve grid resilience as distributed renewable generation expands.
- Robots could gather diagnostic data in hazardous industrial areas without exposing workers to the same risks.
- Within 18 months, AI-assisted coding could let domain experts create industrial applications without traditional software backgrounds.
Risks & concerns
- Almost half the industrial workforce is set to retire within five years, threatening a major loss of operational expertise.
- Less explainable AI models can evolve over time, making their behavior harder to predict around mission-critical physical equipment.
- Automated set-point changes can create safety and reliability risks unless actions remain within tightly controlled operating bands.
- No agreed methodology currently measures AI’s environmental impact across electricity, resources, water and carbon.
- Plants, power systems and mines must redesign processes and validate application-specific guardrails before deploying greater autonomy.