McDermott leverages AI to save valuable engineering hours with user-friendly VIKTOR apps

• 100 P&IDs checked in about 3 minutes
• Cuts review from tens of hrs to mins
• Also built for isometric drawings

McDermott selected VIKTOR.AI to build the ReCo tool because of the platform's focus on engineering rather than generic software development. As Mina Ya-alimadad, AI Projects Manager & Academia Partnership Lead at McDermott, put it: "We chose VIKTOR.AI for our application development because of their focus on engineering. We speak the same language. VIKTOR.AI immediately understood our business needs and the needs of our end-users, who are ultimately our engineers."

The ReCo tool itself was built with usability for process and piping engineers in mind, not just AI specialists, and was developed in collaboration with a Controls and Instrumentation Engineer to match real P&ID review practices. Ya-alimadad noted: "The app is very simple in terms of the user interface, so it's easy to use for any process engineer. Most of the complicated work is done in the background."

Process and piping engineers at McDermott were spending a significant share of their time on manual, repetitive work that added little unique engineering value, checking P&IDs page by page for errors that a trained eye could catch but that took hours to work through on any given project.

McDermott's AI team also faced the broader challenge of introducing AI into an engineering culture that hadn't used it before. Beyond building a working tool, they needed something simple enough for process engineers themselves to adopt directly, without requiring AI expertise, while also demonstrating to the wider engineering organization what AI-based tools could realistically achieve.

Before the ReCo tool, checking Piping and Instrumentation Diagrams for mistakes was a manual process carried out by highly experienced process and piping engineers. Individual P&IDs can run up to 100 pages, and each one had to be reviewed by eye for safety, planning, and cost reasons before a project could move forward.

That manual review is thorough but slow. According to Mina Ya-alimadad, AI Projects Manager at McDermott, these checks "can take up to tens of hours per project," time that senior engineers could otherwise spend on higher-value engineering work rather than repetitive diagram checking.

VIKTOR.AI let McDermott's AI team turn the ReCo tool concept into a working, engineer-friendly application that automates a task engineers previously did by hand. The tool uses image recognition to process P&IDs and flag likely mistakes, cutting review work that could take up to tens of hours per project down to just a few minutes; the tool processes 100 P&IDs in about 3 minutes.

Because the interface was built to be simple enough for any process engineer to use directly, the time savings did not come with a steep learning curve or a need for AI specialists to run the tool day to day.

The project's impact goes beyond a single tool. As Ya-alimadad put it, "With the ReCo tool we introduce AI to the engineering team and show them what can be achieved. We are not just building tools. We are educating people on AI and demonstrate the value it has for their work." McDermott has since built a companion ReCo tool for isometric drawings, and uses the original tool as a training aid for junior engineers.