Understanding Best Practices in Applying AI for Autonomous Operations
Panel 4 brought a measured and experience led perspective to the discussion around autonomous operations, emphasising that progress in this area is evolutionary rather than revolutionary. Drawing on decades of operational, control, and enterprise architecture experience, panellists stressed that many of the foundations for autonomy already exist in the form of established control systems, analytics, and physics based models, with AI extending these capabilities rather than replacing them.
A central theme was risk proportionality. Speakers highlighted the need to align levels of autonomy with operational criticality, arguing that deterministic, explainable approaches remain essential close to the control room and physical assets. Generative and agent based technologies were positioned as more appropriate further up the value chain, where contextualisation, workflow automation, and decision support can deliver value without introducing unacceptable risk.
Trust, explainability, and governance emerged as prerequisites for scale. The panel agreed that autonomous decision making must be auditable, transparent, and bounded by clear escalation paths to human operators. Without these guardrails, organisations risk undermining confidence in systems that are intended to improve safety and performance. The importance of outcome led use case selection was repeatedly reinforced, with panellists cautioning against deploying AI simply because it is available.
Looking ahead, the discussion explored how advances in compute, edge processing, connectivity, and model lifecycle management will gradually expand the scope of autonomy. However, panellists were clear that legacy infrastructure, organisational readiness, and data contextualisation remain defining constraints. The consensus was that meaningful progress will come from augmenting human decision making today, building trust incrementally, and designing systems that can mature responsibly over time.
Speakers
- Amit Jain , Upstream & Downstream, Engineering & Operations, Chevron
- Gerhard Koch , Vice President Digital Products & Solutions, Siemens Energ
- Prashant Jagarlapudi , Sr. Director of Portfolio for the Data and Analytics pillar, Honeywell Connected Industrials
- Varun Anand , AI Strategy | Enterprise Architecture | Innovation Leadership
Moderator
