Oil & Gas companies are facing increasing pressure deliver a new energy future marked by improved efficiency, increased safety, and reduced environmental impact. To deliver on these imperatives, industry leaders like Royal Dutch Shell have launched initiatives to embed AI and ML in their core business practices.

These initiatives are led by digital executives working with teams of developers and data scientists to deliver high-value applications. But front-line analysts, if armed with the right tools, can also contribute predictive solutions to everyday challenges and accelerate O&G’s transformation.

Two Kinds of Digital Initiatives

The AI projects that garner the most press are enterprise efforts—high-value, wide-reaching, and technically complex. Take for example, Shell’s deployment of a predictive maintenance application across hundreds of thousands of valves around the globe. This application integrates data from dozens of enterprise systems datastores, runs and maintains millions of ML models in production, and delivers the right predictions to the right engineers in time to forestall failure. It touches dozens of business and engineering processes, hundreds of end users, and terabytes of data.

But the majority of O&G ML use cases do not reach this scale. Engineers across upstream, midstream, and downstream operations, crude sellers and traders, logisticians and supply chain managers, and retailers work day in and day out with large datasets on complex analytical problems, many of which would benefit from predictive approaches. Take a process engineer seeking to optimize output quality for an underperforming process by predicting key variables based on setpoint and lab test data. Or a trader interested in ad hoc demand forecasting using historical deal and real-time market data.

Companies don’t have the resources to staff ad hoc use cases like these with data scientists and developers. But if front-line engineers and analysts are given products capable of delivering predictive solutions on unlimited datasets, they can drive substantial improvements in productivity and efficiency.

Front-line Requirements

To deliver real value to front-line engineers and analysts, such a product must deliver the following:

  • End-to-end capabilities. The product must be capable of integrating any relevant data—from sensors, asset frameworks, and operational datastores in the case of the reliability or process engineer, and from market, pricing, and customer systems in the case of the trader. It must enable rapid cleanup, merging, and filtering on this data so that analysts can spot trends down to the asset and customer level. It must provide comprehensive capabilities for building predictive ML models. And it must enable users to explore the insights generated within the product or send them to other applications to be acted on. Without these end-to-end capabilities, engineers and analysts spend more time jumping between systems than they do addressing actual use cases.
  • Intuitive, no-code experience. Code or low-code products, no-code products with user interfaces built in the 90’s, and products with months-long learning curves simply aren’t worth the time and effort for busy front-line analysts. A useful product will deliver an intuitive, self-service experience—complete user support in the form of documentation and a robust user community, and guardrails and explainability packages to guide any machine learning workflow.
  • Proven industry success. Front-line engineers and analysts don’t have time to waste on grand analytical experiments. They need to work in products that have proved their worth by solving similar use cases for other analysts in the O&G community.
BHC3 Ex Machina

We are excited to announce the general availability of BHC3 Ex Machina—a no-code software product designed to increase O&G engineer and analyst productivity and efficiency with comprehensive, end-to-end capabilities for data ingest, analysis, and exploration; ML model development and interpretation; and visualization and reporting. BHC3 Ex Machina allows reliability engineers, process engineers, systems & instrumentation engineers, and marketers and traders to deliver predictive solutions for their common analytics challenges:

  • No-code ML enables users to address anomaly detection & predictive analytics use cases without deep data science expertise
  • Native timeseries capabilities enable intuitive, rapid exploration and analysis of sensor data
  • Data ingest, prep, and blend features enable rapid, easy, repeatable analyses & reporting on data from any O&G datastore
  • Intuitive no-code interface speeds up analytical work from data access through acting on insights
  • And finally, users benefit from a product with proven success in O&G. Read how others in the industry have benefited from the combined expertise of BakerHughes and C3.ai by addressing challenges ranging from process optimization to bad actor part identification to sales and marketing optimization here.

Visit the website to learn more about the product and start your free trial today.