Honeywell's CEO says AI's big gain won't come from productivity

As CEO of one of the world's largest industrial groups HoneywellVimal Kapur doesn't think about AI like most people.

It's not about the threatened office worker. “There is always a trend that makes your skills obsolete every five years,” Kapur said recently at the CNBC Evolve: AI Opportunity Summit in New York City. “The exodus of employees is a continuous development.”

And he said it's not about the cool features that might be offered to the consumer who is “excited about writing a resume or a restaurant recommendation.”

The biggest problems AI can solve at Honeywell start with a generational workforce shortage facing the company and its customer companies. From pilots to technicians, declining birth rates in the industrialized world have left fewer people available for jobs that were popular 25 years ago. “Everyone in the industry has this problem,” he said.

The AI ​​opportunity for Honeywell is to create a new workforce pool that can learn and work alongside AI and accumulate and deploy institutional knowledge much more quickly. He said the 15 years of experience traditionally required for a human to handle a complex role can be achieved at the same level by someone with five years of experience working with two AI co-pilots.

Workforce is not the only issue where AI is being used. Kapur pointed to Honeywell's planned launch of engine connectivity in the next few months, which will allow the company to proactively monitor engine performance for maintenance issues before engines return to the shop floor. The same goes for smoke detectors, another Honeywell staple, which are identified for service or replacement much sooner than before.

But it's the labor issue that remains top of mind for Honeywell's CEO, and he added that it's leading him to view AI as a way to generate revenue rather than a productivity solution. “The shortage of skilled workers is the core of the problem for us,” said Kapur. “It is a hindrance to sales growth. The biggest obstacle to sales is the lack of qualified workers.”

Most companies are just beginning to explore the return on investment in AI at a level far removed from the underlying large language models of OpenAI and Nvidia's chip manufacturing.

Jake Loosararian, CEO of Gecko Robotics, whose company works in energy, manufacturing and defense to optimize maintenance efforts – his AI-powered inspection robots analyze equipment as large as aircraft carriers to identify structural defects – says the raw data that Directly from the source without filtering through intermediaries will be the key to the AI ​​success of many companies.

“The future belongs to companies with first-order data sets,” he told CNBC “Closing Bell Overtime” host Jon Fortt at the Evolve: AI Opportunity event.

The importance of moving beyond the current focus on the large language models was emphasized by several executives, including Clément Delangue, co-founder and CEO of Hugging Face, one of the world's most highly valued AI startups, working at the forefront of LLMs World, powered by Amazon, Nvidia and Google. At the CNBC event, he expressed a similar opinion to Loosararian.

“Data and datasets are the next frontier for AI,” said Delangue. He noted that over 200,000 public datasets have been shared on Hugging Face's platform, which uses an open source approach to developing AI models, and the growth rate of datasets added to the platform is faster than the growth rate of new large language models.

“The world will evolve to where every single company, every single industry, even every single use case will have its own specific, tailored models,” Delangue said. “Ultimately, just as every company has its own code repository and creates its own software products, it will create its own models… and ultimately that will help it differentiate itself.”

As companies get the most value from AI tailored to their use cases, this is coupled with a view that is becoming more prominent in AI regulation discussions, shifting the focus away from large language models and towards industry-specific monitoring. And as these use cases become more common, the C-suite must ensure they are communicated to the board.

“Board members need to really understand what the use cases might be for their company so that they can get the report from the people who are most knowledgeable about the risks their company may face,” said Katherine Forrest, a former federal judge and Partner at law firm Paul, Weiss, Rifkind, Wharton & Garrison, an AI legal expert, said at the CNBC AI Summit.

She said now is the time to ask: “What are risks? Do we have the right people managing these risks? Have there been any incidents? They should be aware of the actual realization of these risks.”

For all the debate about how quickly AI opportunities will emerge, Honeywell's Kapur is optimistic that the adoption curve will steepen quickly. “Awareness is high, acceptance is low, but there will be a tipping point,” he said. “I really believe 2025-2026 will be a big year for AI adoption in industry.”