A major oil and gas operator has completed a major overhaul of its engineering requirements management process, leveraging artificial intelligence (AI) to streamline document handling, tagging and rewriting — cutting project turnaround time from months to weeks.
The company, which was facing growing inefficiencies in processing highly complex technical standards, initiated the program as part of a wider digital transformation strategy aimed at improving precision, reducing costs and boosting productivity.
“Our goal is to create a more cohesive, user-friendly and digitalised requirements library that enables our capital project teams to deliver projects with greater precision and effectiveness,” the client said.
The challenge was significant. The operator’s internal repository contained more than 750 documents, each running to 30 pages with over 100 interlinked requirements.
Traditional manual approaches to aligning these documents with industry standards were proving too slow and error-prone.
Meanwhile, engineers could only commit a fraction of their time to the effort, which meant throughput remained low.
The company began by standardising its engineering requirements against international frameworks, including the International Council on Systems Engineering (INCOSE) and the Easy Approach to Requirements Syntax (EARS).
Each requirement also needed metadata tags selected from a library of over 1,000 options — a task that had previously consumed extensive manpower.
AI was identified as a potential solution to automatically rewrite documents and apply metadata, transforming unstructured requirements into structured, searchable data.
A proof-of-concept platform was developed in a secure environment, using mock data to demonstrate the technology’s accuracy.
Once confidence was established, the model was applied to actual engineering standards, with subject-matter experts validating results.
The client remarked: “This solution positions us to transform and integrate our current repository of unstructured data into more structured databases, enabling us to leverage AI capabilities in our business processes, such as in design, procurement or in operations.”
Additional features introduced during the pilot phase allowed the system to identify similar requirements, build cross-references, and manage document imports and exports.
Using a combination of Generative AI, Machine Learning, and vector database storage, documents could be processed far more quickly and consistently than before.
The transformation cut the time needed to handle requirements from handling just four or five documents per month to processing 750 in only three weeks.
The company estimates more than 90 per cent cost savings compared to manual methods, with productivity gains worth up to US$5 million over two years.
Beyond speed and savings, the initiative has given the engineering function the ability to repurpose the technology for other high-value uses such as contracting, inspection planning and safety compliance.
But while the success story demonstrates AI’s power to improve efficiency in highly technical industries, some observers note that heavy reliance on generative models raises new questions.
The question of whether AI truly has the power to refine oil and gas efficiency without compromising safety and accuracy still persists.
As more engineering workflows embrace automation, ensuring the right balance between human oversight and machine decision-making may be one of the sector’s defining challenges in the years ahead.
