Tema’s COO explained how digitalization and AI can improve safety in the oil and gas industry.

Innovarpel 2025, the technical conference on Industrial Digitalization in the Oil & Gas industry, organized by Arpel (Association of Oil, Gas and Renewable Energy Companies of Latin America and the Caribbean), took place in Rio de Janeiro (Brazil) on June 24 and 25.

The program of this edition, which coincided with Arpel’s 60th anniversary, was attended by more than 50 experts and the presentation of more than twenty technical implementation cases. One of the case studies was shared by Albert Tasias Francí, Director of Operations at Tema, who spoke about how digitalization and Generative Artificial Intelligence can transform management in process safety decision making.

In his presentation, he described how one operator took responsibility for a mature field and, after running eleven risk studies (HAZOP, What-If, HAZID, LOPA, QRA and FERA), came up with 404 safety recommendations to be implemented. He also explained how, faced with this large volume of information, the digital management platform made it possible to group and reduce these 404 measures to just 52, and then select only 13 risk reduction measures. This solution made it possible to trace each decision, facilitating collaboration between teams and documenting the entire cycle, from analysis to execution. Attendees were able to see how, thanks to digitization and the centralization of criteria and workflows, it is possible to maintain the chain of custody of each recommendation and speed up decision-making.

In parallel, Tasias also evaluated the use of generative artificial intelligence to assist in three key phases: classification and clustering of security recommendations, data collection to perform ALARP reasoning, and final prioritization.

These would be the main conclusions of the presentation:

• A clear and structured process is needed to move from risk studies to actions.
• Digitization is essential to ensure traceability, collaboration and better decisions.
• Generative AI has potential. It is useful for gathering data and supporting junior engineers, but it is not yet a replacement for senior experts and requires human supervision and more training before large-scale deployment. As Tasias noted, “it is an ally under construction, with room for improvement ahead.”