This story is part of Big Thinkers, a series of profiles on business leaders transforming industries with bold ideas.

“If someone shuts the window, go to the door,” Nancy Greco told Industrious from her research lab in Yorktown Heights, New York. “And if the door doesn’t open, blow a hole through the wall. Don’t let anything be a barrier to you moving ahead.”

Greco is an IBM Distinguished Engineer, a role focused on building the boldest innovations in areas like programming, services, science, design and technology.

Her mantra: be curious and make time to understand how and why things work, even if that means circumventing the status quo. She’s had a long career with IBM—starting in 1981—but has never allowed company life to pin her to one role.

“You can’t meet me and figure out what I am,” she said. “It takes a little while, because I’m not just one thing. I am a compilation of many roles, and many lives.”

See—and hear—solutions for building a smarter factory

Greco’s life has never been narrow. She grew up on a farm in the Hudson Valley, where her family raised dairy cows. Relatives called her an “odd duck” for wanting to read rather than play outside. Since farm life was her primary world, it’s no surprise that Greco initially wanted to be a veterinarian and would often ask her family questions about why some animals got sick while others stayed healthy. She also observed how her family researched symptoms, treatments and preventative measures for their animals.

Similar challenges inspire Greco’s work with next-generation computer systems today. When an automotive part needs a massive recall due to a bad weld, for example, she considers that a form of sickness.

In the manufacturing world, inspections are typically performed visually or sometimes destructively—meaning that the piece is physically destroyed to inspect whether it was properly made. Every second counts in the fast-paced manufacturing world. If it takes minutes to detect a bad weld, multiple products may have been manufactured with that defective weld.

Greco and her team have become resourceful when it comes to inspections. In one of their projects, they’ve included sound as an added inspection step for humans working with AI.

“With acoustics, AI can tell within seconds that a weld is bad and stop the robot,” Greco said. “The human is needed to collect, teach and validate the AI model.”

And now instead of inspecting for the defect, the human can spend time on troubleshooting and creating preventive measures.

Thanks to Greco and team’s innovation, manufacturers from automotive to electronics are responding positively to this approach.

Greco attributes much of her success to her two contrasting degrees. As she self-funded her way through Cornell University, she first decided to study chemistry. That often placed her in an introspective setting where she didn’t need to speak a lot—which as an introvert, she welcomed. But since she also faces fears head-on, she took on an unexpected challenge, and pushed herself to become a college radio DJ.

“I said to myself, ‘Get in front of a mic,’” Greco said. “’How hard could it be?’”

While scary at first, Greco soon began looking forward to seeing the red light come on, her cue to fill the silence with words people would want to listen to.

“You’ll notice a lot of the DJs have a certain voice,” she said.

She demonstrates by dropping her voice to a lower register: “‘Welcome to WBNR.’ I started to learn techniques like that. I started to get exposed to speaking style.”

For Greco, learning how to talk to an audience led to her second degree in technical communication.

“If people don’t understand what you’re saying, they don’t get it, they don’t buy it, and they don’t use it,” she said.

The ability to both understand and ex- plain the most complex inventions—she holds more than 20 patents—makes Greco a captivating storyteller. When she talks about building semi-conductors, for instance, she explains it as “building a micro-city, layer by layer.”

An example of Greco’s approach is her work with data. Around 80 percent of the world’s data is proprietary and not shared outside of its organization. Greco works with these entities to help them realize the potential their data holds, and how AI can release the insight from the data, and then how to share the AI models, not the data, to enable faster learning.

With manufacturers’ data, for example, consistency is key in identifying defects or problems. AI delivers that consistency, providing hour after hour of inspection.

And AI not only takes on burdensome tasks like inspection, it adds context to the data so the human can understand quickly what actions need to be taken. As an example: AI identifies a dent in a part, and also adds where the dent was likely to occur, what tool was used and why.

“Let’s release the human to do the creativity, to do the problem solving, to create new ways of doing things,” she said. “There’s so much more that humans could do with their brainpower.”