Did you miss a session from the Future of Work Summit? Head over to our Future of Work Summit on-demand library to stream.
Building artificial intelligence into your products, services, and processes can make you smarter, faster, and better able to compete. But building smart systems using machine learning is not like buying an accounting package or an enterprise resource planning system.
That’s why executives need as much training as engineers when adopting AI, said Larry Pizette, the head of of data science at Amazon’s Machine Learning Solutions Lab, in the latest edition of The AI Show from VentureBeat. It’s also key to understanding the major mistakes companies make when they’re kicking off AI projects.
“The part that I think gets missed frequently is teaching the business folks, because people always think about the data scientists and the software developers learning about these skills,” said Pizette. “The business folks need to learn as well.”
As part of his role leading data science for Amazon’s machine learning solutions group, Pizette has worked with over 175 companies on AI projects. That experience has taught him a thing or two about what works and what doesn’t.
Executives are used to purchasing software systems, but AI is not like traditional systems. Rather than purchasing a static solution that does job A or B, a machine learning component of a business strategy is more like purchasing a process or a way of thinking about a business challenge. And it requires ongoing input, tuning, and training.
“With machine learning, it’s a little bit different for business owners,” said Pizette. “Let’s say you’re predicting home purchases, but interest rates change. If your model is trained on some assumptions and now something changes in the future, you have to retrain your model. So training the business folks so they understand what they’re getting into and how to best procure it, I think is super important.”
Many businesses know they urgently need to adopt machine learning, Pizette says, but don’t understanding what that means from a process perspective.
The biggest mistake companies make? Not surprisingly, since we know that AI is data-hungry, it has to do with data — but less obviously, according to Pizette, it’s about the people who own the data.
“So the mistakes that people make are typically more around the human element than the technical element,” Pizette said. “[This is especially about] not having the data people in the room when they’re thinking through what they want to do.”
Business owners can have a vision, but without the data to support that, any machine learning projects will be starved for input. So having your data analysts, scientists, and administrators present is essential. They can also fill in gaps, Pizette says. Pretty much everyone’s data is incomplete or has quality issues, but data scientists can work through these and ensure you have enough clean data to get started.
Another mistake businesses make is thinking too far ahead. Having a long-range vision is important, but building out a massive multi-year strategy is asking for trouble. AI and machine learning systems are built to grow. It’s almost impossible to know up front where that might take you over time, so spending weeks and months on long-term planning is overkill. Possibly worse, it often leads to analysis paralysis.
“I’ve seen some organizations want to do so much planning that it keeps them from getting going,” Pizette said.
Another major challenge may also be one of the key reasons business executives need training as much as developers: It’s not always easy to understand the way machine learning works. Looking under the hood might not sound fun to a finance executive or a CEO, but it’s critical to being able to assess both the investment required and the payback potential for an AI initiative.
Machine learning is different, Pizette says, from the rule-based systems most corporate leaders are accustomed to using and purchasing.
“With machine learning, it’s a little bit different for business owners to say, ‘I am going to be acquiring a system that’s making predictions, and how do those predictions affect my business?'” Pizette says. “‘And then what happens if those predictions stop being as accurate as I need them to be?’… If your model is trained on some assumptions and now something changes in the future, you have to retrain your model. So training the business folks so they understand what they’re getting into … is super important.”
To hear the full conversation, watch the video above, or subscribe to the podcast on your platform of choice. Pizette dives deep into training for developers, projects the’s worked on for clients like the NFL, and where we are now in terms of AI development.
VentureBeatVentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the subjects of interest to you
- our newsletters
- gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
- networking features, and more