Trust is key for the wider adoption of AI in mission-critical industries, says Cyient CTO
With Artificial Intelligence (AI) a hot topic at this year’s Farnborough International Airshow, few companies understand what’s at stake as well as Cyient. Karapattu…
With Artificial Intelligence (AI) a hot topic at this year’s Farnborough International Airshow, few companies understand what’s at stake as well as Cyient. Karapattu Arunachalam Prabhakaran explains why trust is the real barrier to adoption, not the technology itself.
Cyient has a 35-year legacy in mission-critical industries, including aerospace, energy and mining sectors, building what Karapattu Arunachalam Prabhakaran, Senior Vice President & Chief Technology Officer, describes as deep domain and process knowledge over more than three decades. That experience, he says, is now being channelled into a new challenge – “building AI solutions that customers can trust”.
Cyient explains why AI adoption is stalling in critical industries
“We hear plenty of hype about AI in the market,” says Prabhakaran. “AI might be good at software development in a non-critical commercial environment, but aerospace and other mission-critical sectors are a different ball game,” he adds. They are defined by extreme complexity and zero tolerance for error. “AI can’t be right most of the time. It has to be right all the time.”
Prabhakaran identifies three barriers holding the sector back. The first is the data lifecycle. The data has to be connected from engine design through to maintenance and operations. It can’t be isolated. The second challenge is a tendency to focus on technical proofs of concept rather than on a clear business outcome, such as reducing product development time or maintenance costs. The third, and most significant, is trust. “Today, mission-critical engineers don’t trust AI recommendations, [as they are] not being right all the time.”
Cyient’s approach, explains Prabhakaran, begins with connecting data across the value chain and giving AI the context it needs to make sound recommendations. “For AI to make a recommendation, it needs to understand the design background. It needs to understand the maintenance approach and the operations data. Then it can understand the concept and make a sound recommendation.”
That level of thinking underpins three lifecycle solutions being developed by Cyient. “AI-enabled engineering, AI in the aftermarket and AI in quality and regulatory compliance.” It’s not about more data, reiterates Prabhakaran. “It’s about the right data.”
Trust has to be built in from day one
For Prabhakaran, intelligence cannot simply be layered on after a system is deployed. “It has to be engineered from day one, particularly for mission-critical industries.” That means building explainability, certification constraints and architectural governance into the system’s architecture from the outset, connected through a digital thread that ties design, engineering and operations data together in a meaningful way. “That builds the trust and credibility for AI solutions.” Above all, he stresses, “AI projects must start from the right business outcome, so it can drive that success.”
AI will complement engineers, not replace them
Referencing concerns that AI could displace the human workforce, Prabhakaran is adamant that won’t be the case in mission-critical industries. “AI will not replace. It will augment the domain knowledge, the judgment and the accountability that human engineers bring in,” he says, describing the shift as “collaborative engineering” powered by AI’s computational strength alongside human expertise.
That collaboration matters more than ever, he adds, as experienced engineers retire, often taking that institutional knowledge with them. “That’s where AI could really complement and capture that human knowledge, while preserving that expertise. “For mission-critical industries, that’s going to be the future.”
