• Reviewing real-world results from predictive maintenance, production optimization, and leak detection to demonstrate tangible improvements in uptime, output, and emissions.
• Comparing pilot project results with year-round operational data to highlight the real-world impact and sustainability of AI solutions.
• Analyzing why some projects stalled-due to poor data, IT/OT handoff issues, or model drift-to help attendees avoid common and costly pitfalls.
• Identifying scalable patterns like small on-site models and trusted data search to guide effective and expandable AI deployment across assets.

• Defining clear boundaries for what AI can see, suggest, and automatically change to reduce operational risk and prevent scope creep.
• Establishing approval workflows and emergency stop procedures to ensure human accountability remains central to safety-critical decisions.
• Testing all AI-driven changes in a safe, simulated environment before live deployment to prove reliability and prevent costly downtime.
• Recording every AI action and human override to create a transparent audit trail for compliance, incident reviews, and continuous improvement.

• Sharing quantified operator results for predictive maintenance, production tuning, and leak detection to identify high-value, repeatable wins.
• Translating AI outcomes into business metrics like uptime and throughput to build compelling business cases and secure executive funding.
• Detailing the playbook for moving from a successful pilot to a multi-site rollout to accelerate deployment and maximize ROI.
• Highlighting common pitfalls in data, change management, and oversight to protect safety and ensure long-term user adoption.

Three rapid 10-minute talks to spark discovery. One problem, one approach, one result.

• Implementing RAG (Retrieval-Augmented Generation) with effective connectors, caching, and guardrails to ensure AI provides accurate, context-aware answers from enterprise data. - Gulshan Singh, Global Data Analytics Manager, Huntsman
• Structuring schemas and metadata to dramatically improve the quality and relevance of information retrieved by AI systems - John Green. Analyst, Kinetik

• Bridging IT, OT, and business teams to align digital and AI initiatives with real plant challenges improving reliability, throughput, and workforce efficiency.
• Building a centralized digital platformthat integrates SAP, historians, and cloud services into a unified layer for analytics, AI, and decision support.
• Applying practical AI and machine learning for equipment health and anomaly detection, keeping operators in the loop with clear red–yellow–green guidance.
• Accelerating workforce readiness by capturing expert knowledge, guiding new engineers in real time, and designing user experiences for the next generation.

• Leveraging the OSDU data platform to break down subsurface and operations data silos, creating a unified foundation for AI applications.
• Ensuring AI models have access to trusted, well-contextualized data with clear lineage to improve the accuracy and reliability of their outputs.
• Navigating the practical challenges of implementing OSDU from data migration to taxonomy alignment to enable scalable AI solutions across the asset lifecycle.

• Designing approval chains and audit trails that align with existing OT safety procedures to build trust with field personnel and engineers.
• Implementing kill-switches and rollback protocols that give operations ultimate control over any AI-driven action or recommendation.
• Clarifying the RACI matrix for AI sign-off in OT environments to ensure clear accountability and prevent operational disruptions.

• Security teams in remote operations make critical decisions with incomplete information, conflicting local reports, and verification delays that can stretch from hours to days.
• Discover how AI fused radio monitoring, sparse social media, and movement patterns in Chad into actionable forecasts, reducing decision latency while field teams maintained operational schedules.
• Learn what the model missed, when field operators overrode AI forecasts based on ground truth, and how the team calibrated trust between algorithmic predictions and local knowledge.
• Understand deployment in Chad, an austere technical environment: intermittent power and bandwidth, cloud access limitations, and strategies for handling potential data manipulation.
• See how to make actionable a proven approach that applies to emerging basins, offshore logistics, and other operations where information scarcity creates risk but operational continuity is essential.

Quantum Tech, celebrating its 100 year anniversary in 2025, is the promise of new computational capabilities, more powerful AI and transformative use cases for the energy sector. In this fireside chat some of the industry pioneers building these technologies will reveal the curtain on the truth - is quantum tech ready? How can you use it? What are the real applications in the energy sector? And why you must get started now even if there is no imminent ROI.

Join some of the most recognized founders and scientists from the quantum space for this engaging conversation that will leave you spooked and excited, with practical tips on how and where to get started with quantum now without large commitments or investment.

· Evaluating the trade-offs between large language models (LLMs) and smaller, domain-specific models based on cost, latency, and accuracy requirements.

· Selecting the right model architecture for the specific task from generative text applications to real-time sensor analytics to optimize performance and value.

· Balancing the flexibility of general-purpose LLMs with the precision and efficiency of smaller models to meet the stringent demands of operational environments.

Rapid 10-minute talks to spark discovery. One problem, one approach, one result.

• Designing approval and escalation workflows for autonomous AI agents to ensure humans are involved in critical or exceptional decisions.
• Selecting the right AI tool for each task based on a balance of cost, latency, and accuracy to optimize overall system performance and ROI. - Meenakshi Mishra, Principal Data Scientist, ExxonMobil
• Determining the optimal points for human-in-the-loop interaction to maintain safety and control without creating unnecessary bottlenecks.

• Standardizing asset data models across sites to build a scalable digital twin foundation on top of PI Asset Framework and the Covestro Monitoring Platform.
• Detecting early anomalies on heat exchangers, safety valves and other critical equipment using rule based and ML based monitoring to avoid unplanned outages and safety risks.
• Triggering targeted maintenance actions through automated alerts and intuitive dashboards so engineers can act before fouling, leaks or degradation impact production.
• Sustaining value by defining clear workflows for reviewing alerts, tuning models and reducing false positives so digital twins remain trusted tools for reliability teams

Get a rapid-fire preview of the future in this high-energy session. Ten hand-picked start-ups will pitch their groundbreaking AI solutions for oil and gas in just six minutes each.

• Witnessing a curated showcase of the most promising AI technologies, from sector-specific LLMs to robust edge computing platforms.
• Evaluating real-world applications and quantifiable results that address your toughest operational challenges.
• Voting live for the most compelling solution, crowning the winner and shaping the future of the vendor landscape.
• Connecting with the next generation of technology partners poised to transform the industry.

o MIN 0-1: FIND your most compelling takeaway from the Lightning Round.

o MIN 1-3: SHARE with a partner: The idea is… and it’s valuable because…

o MIN 3-4: COMMIT to one next step: “One thing I will do to explore this is…

o MIN 4-5: SUBMIT your partner’s most promising action step via Slido.

• Building a strategic portfolio approach to AI investment to move from scattered pilots to coordinated, enterprise-wide transformation.
• Establishing a center of excellence and cross-functional steering groups to standardize best practices and accelerate organization-wide learning.
• Developing a vendor and technology strategy that balances build-vs-buy decisions to ensure scalability, interoperability, and long-term value.

• Defining the clear thresholds and business rules that determine when an AI agent can act autonomously and when it must seek human approval.
• Implementing robust audit trails and rollback capabilities for all autonomous actions to ensure safety and provide accountability.
• Balancing the efficiency gains of automation with the irreplaceable judgment of human experts in complex or high-consequence situations.

• Applying non-deterministic AI models to real-time process data to identify hidden optimization opportunities for energy consumption on brownfield assets.
• Reducing energy costs and carbon emissions without major capital investment by making continuous, AI-driven adjustments to existing control systems.
• Demonstrating how AI can learn and adapt to the unique operating signature of individual assets to uncover sustained performance gains.

• Leveraging AI to analyze LDAR and sensor data, dramatically reducing false positives to focus field crews on genuine leak events.
• Accelerating the find-to-fix timeline by providing crews with precise location data and prioritized repair schedules to maximize emissions reduction.
• Improving the overall efficiency and cost-effectiveness of methane management programs to meet regulatory and ESG commitments.

• Session by Dataiku
• Implementing exception-based surveillance powered by AI to shift engineers from monitoring data to managing by exception, focusing only on assets needing attention.
• Generating daily AI summaries for well and asset performance triage to accelerate decision-making and prioritize intervention efforts.
• Using machine learning to identify schedule risks in turnaround planning, enabling proactive mitigation and protecting project timelines.

• Codifying the specific scenarios, risk levels, and data confidence thresholds that trigger a mandatory hand-off from AI to a human operator.
• Designing seamless user interfaces that make hand-off requests clear, context-rich, and easy for humans to action quickly.
• Ensuring these rules are baked into the core of AI system design to build a foundational layer of operational safety and trust.

Move from presentation to problem-solving in these highly interactive, peer-driven roundtables. This is your opportunity to roll up your sleeves and tackle your most pressing AI implementation challenges with a small group of fellow operators facing the same hurdles.

Choose one topic that aligns with your most urgent priority.

• Building Your AI Center of Excellence: Scope, Charter, and RACI
• Proving AI's Bottom-Line Impact: From Unit Economics to Portfolio ROI
• Fueling GenAI with OSDU: Taxonomies, Security, and Speed
• Governing Autonomous Agents: Approval Matrices, Audit Trails, and Rollback
• Winning Hearts and Minds: Change Management for AI on the Frontlines - Mathias Klinkby, Noble Corp 
• The Data Defects That Derail AI: Prioritizing Quality for Trusted Answers
• Your AI Vendor Blueprint: Build, Buy, or Bolt-On for 2026-2028
• LLMs vs. Small Models: Matching the Tool to the Operational Task
• AI at the Edge: Solving Power, Connectivity, and Harsh Environment Challenges
• The AI-Powered M&A Playbook: Rapid Data Integration in the First 12 Weeks

• Business case that clearly articulates the problem and the proposed solution

• Compelling financial metrics that resonate with decision-makers

• Clear risk mitigation and value creation metrics

• Navigating the customer's budget approval process

• Enabling champions to become effective advocates

• Deploying AI models capable of operating reliably in harsh environments with limited power, connectivity, and computing resources.
• Managing model drift at the edge by establishing triggers for retraining and developing efficient data syncing strategies with the cloud.
• Ensuring the robustness and failure-resilience of edge AI systems to maintain operational integrity and data continuity in remote locations.

• Running sophisticated visual inspection and anomaly detection models directly on edge devices to identify leaks and corrosion in real-time.
• Minimizing bandwidth usage and latency by processing video and sensor data locally, sending only alerts and key metadata to central systems.
• Enabling immediate response to integrity threats by providing field personnel with instant notifications and evidence, enhancing safety and reducing environmental impact.

• Building retrieval systems that work in real-world OT environments by focusing on robust connectors, intelligent caching, and safety guardrails.
• Curating metadata and taxonomies that are purpose-built for operations to dramatically improve the relevance and accuracy of AI-generated answers.
• Monitoring data-quality KPIs that are explicitly tied to financial outcomes to focus data management efforts on what truly impacts the bottom line.

• Automating the triage and resolution of common, low-risk service tickets using AI to free up skilled personnel for more complex tasks.
• Learning from historical ticket data to intelligently route and escalate complex issues to the appropriate human expert along with relevant context.
• Reducing resolution times and improving workforce productivity by creating a seamless collaboration between AI and human teams.

Instructions

• MIN 0-1: FIND your most compelling takeaway from the Lightning Round.
• MIN 1-3: SHARE with a partner: The idea is… and it’s valuable because…
• MIN 3-4: COMMIT to one next step: “One thing I will do to explore this is…
• MIN 4-5: SUBMIT your partner’s most promising action step via Slido.

See AI move from concept to concrete reality. In this rapid-fire showcase, we move beyond slides to live demonstrations that prove the operational and financial value of AI on real-world oil and gas challenges.

• Witnessing five distinct, end-to-end AI solutions operating in a live environment to understand their practical application and limitations.
• Evaluating the integration of AI into core operational workflows from subsurface analysis to regulatory compliance to assess their readiness for your assets.
• Validating vendor claims with live metrics and before/after comparisons that demonstrate clear uplift, accuracy, and speed.
• Identifying the specific technology partners and architectural patterns that can accelerate your own AI deployment plans.