Cegal

The oil and gas industry continues to digitize subsurface data, yet much of its most valuable knowledge remains embedded in reports, well documentation, scanned records, and legacy PDFs. While OSDU® provides a foundation for managing structured subsurface information, creating AI systems that deliver trustworthy answers requires far more than simply making documents searchable. This presentation shares lessons learned from building an enterprise AI solution on OSDU, designed to transform both structured and unstructured subsurface information into trusted, traceable, and actionable answers. The journey began with a document-centric retrieval solution and evolved into a broader platform incorporating retrieval optimisation, evaluation frameworks, access-control enforcement, source traceability, and natural-language access to structured OSDU data.

The presentation explores the practical challenges encountered when deploying AI on imperfect information, including context optimisation through multi-agent orchestration, semantic chunking, hybrid retrieval, prompt optimisation, retrieval evaluation, duplicate and competing report versions, incomplete metadata, entitlement-aware access control, and citation-based source transparency.. Particular attention is given to the role of evaluation in guiding optimisation decisions across the entire pipeline and determining when a system should answer confidently and when it should instead admit uncertainty. A central lesson is that trustworthy AI is achieved through optimisation across every layer of the solution. Improvements came not from a single model or technique, but from continuous refinement of ingestion pipelines, retrieval strategies, search parameters, prompts, evaluation methods, and governance controls.

The presentation also examines the next stage of the solution, where natural-language interfaces are extended beyond document retrieval to support live querying of structured OSDU data. Examples include generating answers directly from OSDU entities, handling real-world schema complexity, and integrating with subsurface tools such as DDMS. Throughout the talk, practical lessons, technical trade-offs, and the realities of enterprise deployment are highlighted with a focus on one question: what does it take to trust an AI answer?

As energy companies modernise their data estates, one challenge remains: unlocking value from legacy Petrel project data. While the OSDU™ Data Platform provides a cloud-native, vendor-agnostic foundation for data access and reuse, Petrel data is often locked inside proprietary project structures that are difficult to search, govern, and integrate beyond the desktop.

This presentation shows how Petrel projects can be transformed into OSDU-ready data through automated extraction, enrichment, and metadata normalisation. Using domain-aware tools such as Blueback Project Tracker and purpose-built extraction pipelines, large volumes of Petrel projects can be scanned, assessed, and prepared for ingestion. The session highlights common challenges, including version inconsistencies, incomplete legacy archives, embedded unstructured content, and missing spatial metadata, and explains how validation, enrichment, and reference-data mapping improve quality and alignment with OSDU domain models.

A central theme is the shift from treating Petrel projects as static containers to viewing them as structured sources of reusable domain data. By breaking projects into independently managed data objects, organisations can enable lineage, governance, cross-project comparison, and more collaborative cloud workflows.

The talk shares practical lessons from real-world deployments, including archive scanning, project prioritisation, ingestion integration, and quality assurance feedback loops. Attendees will leave with a clearer view of how to approach legacy-to-cloud transformation at scale and turn trapped project data into governed, discoverable, cloud-ready assets.

Every part of the energy value chain creates data that drives critical business decisions - from production allocation and custody transfer to transportation, storage, reporting and commercial settlement. Yet many organizations still rely on fragmented systems, manual reconciliation and spreadsheets to manage numbers with significant financial and regulatory impact.

In this session, Cegal will present how EnergyX supports a more modern, trusted and auditable approach to energy data management. Built as a cloud-native platform, EnergyX helps automate data collection, validation and reporting across the value chain, enabling earlier insight into data quality issues, reducing manual handovers and creating one governed version of the truth for operational, commercial and financial teams.

The presentation will show how EnergyX helps energy companies move from chasing numbers to trusting them.

As organizations prepare AI-driven workflows and increasing regulatory demands, the importance of trusted technical data has never been greater. Yet many companies continue to operate across disconnected systems, repositories, and disciplines.

Cegal is developing Cenova, a modular ecosystem aimed at providing complete visibility, governance, and intelligence across technical data environments. But the future of data management cannot be designed in isolation.

In this session, we will present our vision, share our current roadmap, and invite attendees to participate in an open discussion on industry priorities. We want to understand where organizations are investing, what challenges remain unresolved, and which capabilities are most important for the next generation of data platforms. This is not a product presentation. It is an opportunity for the industry to influence the direction of a new data ecosystem and help ensure that future solutions solve real-world problems.

How well do you really understand your technical data estate?

Many organizations struggle to answer fundamental questions about their data: What do we have? Where is it stored? Who owns it? Which data is duplicated, unused or creating risk?

In this interactive workshop, participants will use Cenova Command to investigate a real-world technical data environment and uncover hidden insights. Through guided exercises, attendees will learn how to identify data quality issues, assess storage utilization, improve governance, and establish the visibility needed to support future digital and AI initiatives.

No prior experience with Cenova Command is required. The session is designed to provide practical skills that attendees can apply immediately within their own organizations.

As Petrel environments grow in size and complexity, maintaining visibility and control across projects and subsurface data becomes increasingly challenging. Effective data management requires a clear understanding of where data resides, how it is being used, where duplication exists, and whether projects comply with established standards. This paper demonstrates how Blueback Project Tracker can provide a structured approach to monitoring and managing the Petrel data landscape.

Project Tracker scans Petrel projects to build a centralized inventory of project metadata. This provides data managers with visibility across the environment and enables projects and datasets to be monitored over time. Project rules and notifications can be used to identify inconsistencies and support data governance, while integration with tools such as ArcGIS and Power BI enables further spatial analysis and data visualization. Project comparison and hashcode-based methods enable equivalent or diverging datasets to be identified across Petrel projects. Project lineage can also be visualized to understand relationships between copied projects and identify opportunities for consolidation. A seismic rationalization case study illustrates how duplicated seismic data can be identified across multiple projects and consolidated into a common repository, reducing redundancy and improving data organization.

Beyond monitoring and analysis, Project Tracker supports practical data management activities including project upgrades, moves and deletion, as well as Studio repository monitoring and clean-up workflows. Together, these capabilities provide greater visibility and control across the Petrel environment, supporting improved data quality, storage optimization, data governance and more efficient subsurface data management.

Recent advances in agent-based artificial intelligence are transforming how subsurface professionals interact with enterprise data ecosystems, information models, and technical workflows. Instead of manually discovering, validating, and integrating data across multiple systems, intelligent agents act as context-aware assistants that interpret user intent, reason across distributed data sources, and orchestrate end-to-end data management processes.

This work presents a practical demonstration of an agent-driven approach applied to a subsurface use case using an open field dataset. Starting from a high-level objective, the agent discovers and evaluates available data assets across enterprise repositories, summarizes field and reservoir context, inventories wells and associated datasets, and assesses data completeness, consistency, and quality. The agent integrates information from both internal data platforms and external regulatory and public data services to identify and resolve gaps, including missing interpretation data.

Building on this foundation, the agent generates reproducible workflows for data ingestion, transformation, conditioning, and enrichment within a unified environment aligned with enterprise data models. It demonstrates how heterogeneous data can be harmonized through metadata-driven approaches and interoperable services. As part of the workflow, the agent constructs an analytical pipeline, including a machine learning model to estimate missing subsurface properties.

By interacting with data catalogs, services, orchestration frameworks, and computational platforms, the agent highlights the importance of interoperability and standardization. This work shows how agent-based systems bridge the gap between domain intent and governed, reproducible workflows, improving data accessibility, strengthening data quality, and accelerating integrated digital subsurface workflows.

Data management (DM) routines in Petrel and Studio are traditionally manual, repetitive, and resource-intensive. While automation can alleviate these bottlenecks, a major challenge persists: most traditional DM professionals lack the programming or scripting expertise required to build automated workflows independently.

To bridge this gap, we developed an agentic system utilizing a Model Context Protocol (MCP) server integrated with Cegal Python Tool Pro (PTP) and Petrel. This innovative solution empowers traditional DM users with zero prior coding experience to automate routine day-to-day tasks simply by interacting with an agent bot using plain, natural language prompts.

The system dynamically interprets conversational intent and routes requests down targeted operational paths. Routine queries, like well lookups, data summarizing, or attribute checks, are answered directly from trusted sources. Analytical tasks are handled by orchestrated workflows, while high-risk, data-altering actions such as well and well-associated data loading, quality control (QC) validation, data updates, and pushing directly to Studio repositories follow a strictly guarded path requiring explicit human approval before execution.

Behind this single conversational interface, heterogeneous existing assets including custom Python code, Petrel/Studio workflows, and legacy QC utilities are unified under a simplified, OSDU-informed data model. This approach completely democratizes automation, transforming tedious manual work into fast, consistent, and securely auditable workflows for everyday data managers.

This presentation showcases how the image tools in Blueback Toolbox help geoscientists unlock the value of image-based information from multiple sources by bringing it directly into a Petrel project. Attendees will learn how images extracted from PDF reports and online publications can be efficiently imported, managed, and converted into usable Petrel objects. Through two real-world examples, we will demonstrate how to transform an image into a 2D seismic line and how to generate a surface object from an image, enabling interpreters to seamlessly integrate external content into their subsurface workflows without leaving the Petrel environment.

How well do you really understand your technical data estate?

Many organizations struggle to answer fundamental questions about their data: What do we have? Where is it stored? Who owns it? Which data is duplicated, unused or creating risk?

In this interactive workshop, participants will use Cenova Command to investigate a real-world technical data environment and uncover hidden insights. Through guided exercises, attendees will learn how to identify data quality issues, assess storage utilization, improve governance, and establish the visibility needed to support future digital and AI initiatives.

No prior experience with Cenova Command is required. The session is designed to provide practical skills that attendees can apply immediately within their own organizations.