AI Assistants Accelerate Decision-Making and Efficiency at TotalEnergies' Antwerp Refinery

Refining Intelligence: Transforming Operations with AI

Artificial intelligence (AI) is now a key driver of operational efficiency and competitive advantage in the chemical process industries (CPI). Over 80% of CPI executives surveyed by IBM expect AI to significantly impact their business within the next three years. Companies are already applying AI in research and development (74%), manufacturing (61%), risk management (58%), and forecasting (47%) to cut costs, improve product quality, and speed up innovation.

TotalEnergies, based in Paris, is a leading example of AI in action. At its Antwerp refinery, the company introduced AI-powered knowledge assistants—MILAa and JAFAR—to tackle long-standing challenges in knowledge management and root cause analysis. Developed with Sinequa by ChapsVision, these tools combine natural language processing, machine learning, and intelligent search to unify operational data, minimize downtime, and accelerate decision-making.

Challenges with Operational Data

Before these AI tools were implemented, engineers at TotalEnergies struggled to access critical operational data. Root cause analysis (RCA) reports, maintenance logs, and technical manuals were scattered across isolated systems, stored in various formats, and often only available in one language. This manual, document-heavy process caused three main problems:

• Longer downtime due to slow identification of failure causes.
• Inconsistent decisions since knowledge wasn’t easily shared across teams or locations.
• Repeated incidents leading to costly productivity losses, repairs, and supply chain issues.

For example, multiple pump failures over two years could have been prevented if RCA data had been readily accessible across teams in real time.

AI Assistants Built for Industrial Operations

TotalEnergies partnered with Sinequa, a leader in enterprise intelligent search, to create AI assistants that fit the specific needs of engineers and operators.

MILAa: Learning Smarter from Past Failures

MILAa collects over 1,000 RCA documents from the Antwerp refinery into a centralized, AI-powered knowledge base. Using domain-specific ontologies, it extracts and organizes data like equipment types, failure modes, downtime durations, and remediation effectiveness. Engineers can query MILAa in natural language, cutting down the time spent searching through dense reports. This system streamlines problem-solving and helps standardize responses and preventive measures across TotalEnergies’ global operations.

JAFAR: From Documents to Conversational Insights

JAFAR is a generative AI assistant that transforms static documents into interactive, conversational insights. It extends MILAa’s features by automatically translating technical RCA documents into French, Dutch, German, and English. This is possible through a custom internal dictionary that preserves industry-specific terms and acronyms.

“Automatically translating these documents is no easy task, as they contain technical language specific to our core business,” said Pierre Jallais, lead architect at TotalEnergies. “With Sinequa, we enhanced the base model with our internal dictionary, giving JAFAR the context it needs to process our unique documents accurately.”

JAFAR can segment RCA content, identify failure patterns, and summarize data in a way that is both precise and user-friendly, improving how engineers access and use operational knowledge.

AI’s Broader Impact Across the Chemical Industry

Beyond operational knowledge and equipment uptime, AI supports other critical areas in the chemical sector. In R&D, machine learning helps identify promising molecules, optimize formulas, and predict efficacy—speeding up product development and lowering costs. In supply chain planning, AI cuts forecasting errors by up to 50%, improving raw material procurement and reducing inventory waste.

Predictive forecasting powered by AI enables CPI firms to anticipate price shifts and demand changes more effectively, boosting profitability and supply chain flexibility.

TotalEnergies’ AI initiatives reflect these broader trends by applying advanced techniques to improve equipment reliability, maintenance planning, and operational safety.

For those interested in expanding their AI skills in operations or product development, exploring targeted AI courses by job role can provide practical knowledge and tools.