AI In Refining: From Promise to Performance
How artificial intelligence is moving beyond experimentation to deliver measurable value, improve decision-making and strengthen operational performance in refining.
Part of the Latin American Refining Key Takeaways Report 2026 series, this article draws on insights shared by senior industry leaders at LARTC 2026 in Buenos Aires.
Artificial intelligence in refining is moving beyond the initial excitement and proof-of-concept stage towards a more demanding phase: demonstrating measurable business value and scaling successful applications.
The challenge is no longer simply to identify potential use cases, but to ensure that solutions deliver tangible improvements in profitability, reliability and operational performance. Success must be measured by the value generated, rather than the technical performance of a model alone.
Success must be measured by the value generated, rather than the technical performance of a model alone.
The strongest opportunities currently lie in supporting real-time operational and optimisation decisions. Applications such as assisted control-room operations, product-quality prediction, equipment failure forecasting and early risk detection can help refinery teams anticipate issues and respond more effectively.
These tools are most valuable when they complement operational expertise and strengthen decision-making. At the same time, understanding why pilots fail or do not scale remains important for improving technology selection and investment decisions.
Data quality, availability and the underlying digital architecture remain significant barriers to effective deployment and future scalability.
A disciplined implementation approach is therefore essential: first define the operational problem and use case; then assess whether AI is genuinely required or a simpler solution would suffice; establish the necessary data and architecture; evaluate capital requirements and expected returns; and develop a clear plan for user adoption.
This sequence helps avoid unnecessary complexity and directs investment towards applications with a credible path to value.
The challenge is not simply identifying where AI can be used, but ensuring it delivers measurable operational and commercial value.
User adoption is critical throughout the lifecycle of advanced solutions. Operators, engineers and optimisation teams should be involved from the earliest stages so that tools address real operational challenges and fit the way work is performed.
Their involvement helps build trust in recommendations, improve usability and support continuous refinement after deployment. As AI becomes more integrated into operational decisions, organisations must also consider how human expertise, accountability and decision-making responsibilities will evolve.
Ultimately, the transition from experimentation to scaled value depends on combining the right use cases, robust data and architecture, sound investment decisions and meaningful user adoption.
AI can strengthen operational flexibility, improve anticipation of equipment and process issues, and support better decisions in increasingly complex refining environments.
Its long-term contribution will depend less on the sophistication of individual models than on the ability to integrate them into everyday operations and demonstrate sustained business impact.
AI’s long-term contribution will depend less on the sophistication of individual models than on the ability to integrate them into everyday operations and demonstrate sustained business impact.
Latin American Refining Key Takeaways Report 2026
Explore further insights and perspectives shared by senior industry leaders at LARTC 2026 in the full Latin American Refining Key Takeaways Report 2026.
Up next: LARTC Ask the Experts
The conversation doesn’t end in Buenos Aires.
At LARTC Ask the Experts on 4-5 May 2027 in Rio de Janeiro Brazil, continue exploring the questions, challenges and opportunities shaping Latin America’s refining industry, with insights from experienced industry leaders.
Join Us at LARTC 2027 in Cartagena
The next Latin American Refining Technology Conference (LARTC) will return on September 7-9 2027 in Cartagena, Colombia.


