Looking Beyond the Hype – Lessons from Applications of AI in the Energy Industry

Looking Beyond the Hype – Lessons from Applications of AI in the Energy Industry

Join us for an evening of practical insights into how AI is being applied across the energy industry. This session will cut through the hype and focus on real case studies that demonstrate measurable outcomes, lessons learned, and playbooks you can apply in your own organisation.

Date: Wednesday 29 October

Location: Douglas Hotel, Market Street, Aberdeen

Speakers:

• Glen Milne, Production Systems Manager, Spirit Energy
• Sankesh Sundareshwar, Chief Commercial Officer, Voltquant

Both talks will showcase how AI is being deployed responsibly in high-stakes environments, ensuring accuracy, enabling human-AI collaboration, and delivering results at scale.

Abstract from Sankesh Sundareshwar:

This session will explore how AI is driving transformation in capital projects, using real world examples from the oil and gas sector. Drawing on recent work done with Shell, the talk will show how large language models, intelligent agents, and structured knowledge bases have been used.

It will also share how AI is being applied responsibly by ensuring accuracy, avoiding hallucinations, and embedding trust in high stakes environments.

This session is ideal for those working in digital transformation, engineering, capital projects, and operations who are keen to understand the role of AI beyond the hype.

Key Takeaways:

• Real world impact of AI in capital projects

• Human AI collaboration in decision making

• How to get started with AI in your projects

• Lessons learned from deploying AI at scale

About the Speakers:

Sankesh Sundareshwar is an experienced engineering and technology leader with over 20 years in the oil and gas industry. He has led major capital projects across the globe and played a pivotal role in integrating digital innovation into traditional engineering workflows. While at Shell, he spearheaded the creation of the company’s first AI powered tool for capital projects, transforming how teams access data, ensure decision quality, and reduce delivery risk. Today, as Chief Commercial Officer at Voltquant, he continues to drive the use of AI to solve complex challenges in engineering and energy, combining deep domain knowledge with advanced technology to create lasting impact.

Glen Milne:

Glen Milne is Production Systems Manager at Spirit Energy, leading operator-first digital and AI initiatives across production, emissions, and operational excellence. With 20+ years in energy spanning frontline operations through asset management, he focuses on practical tools that plug into workflows and deliver measurable outcomes. Recent work includes cutting an AI procurement project from ~£300k to ~£20k while achieving 95%+ accuracy and a sub-six-month ROI, alongside reductions in maintenance backlog - all by pairing domain expertise with the right model choices and human-in-the-loop verification. Glen speaks regularly on AI and digital transformation (Future Oil & Gas advisory board/panels, regional tech forums) and is known for cutting through hype with operator-grounded case studies and procurement playbooks. A British Army veteran, he blends disciplined execution with hands-on data and systems experience to turn complex problems into scalable solutions.

Abstract for Glens presentation:

Most “AI” proposals for engineering documentation are expensive, generic, and under-deliver because they’re built for clean text, not messy technical drawings. In this talk I share how we flipped the script at Spirit Energy: same problem statement, radically different outcome. A ~£300k large vendor route vs ~£20k delivered, with ~94% savings, 95%+ system accuracy, and ROI in under 6 months. We’ll cover why traditional OCR struggles on P&IDs (it reads characters, not context), and how visual language models (VLMs) combine computer vision with language understanding to recognise equipment, tags, and relationships (e.g., linking TP1 and TP2 correctly). The core lesson isn’t “buy more AI,” it’s procure smarter: prove it on your data, separate real AI from boilerplate code, phase implementation, keep humans-in-the-loop for engineering verification, and insist on measurable milestones. I’ll share the procurement checklist we used, the human-validation pattern that improved results, and where this approach transfers across assets and industries. If you’re evaluating AI for legacy documents, this session will help you avoid the “AI premium” and buy outcomes - not hype.