It’s hard to think of a more buzzed about — or hotly competitive — technology over the last few years than artificial intelligence. As more companies push for new ways to implement deep learning and data science to reap the benefits of automation, it’s clear that an AI revolution isn’t coming: It’s already here.
If your company or organization has been waiting for practical, viable ways to implement AI in your business, there’s arguably never been a better time than now. Here’s how companies in three industries are using AI to make a tangible difference in their businesses.
AI in e-commerce
As consumers continue to push back against pop-up advertising and “creepy” ad tracking, retailers are looking to AI as the key to creating curated product recommendations and improved search experiences in ways that feel natural, not invasive.
“AI helps you actually know who your customers are and what part of the journey they’re in, so you can give them exactly what they’re looking for at the exact right moment,” said Kamelia Aryafar, chief algorithms officer at Overstock, at a recent tech conference in Salt Lake City, UT.
Companies on the forefront of this emerging technology are using AI and deep learning to improve image search and “You Might Also Like” recommendations to both predict what a customer will like but also help them expand their palate. And it’s not just the front-end experience that’s being improved: AI is also making inroads in e-commerce for fraud detection and protection, inventory management, supply chain optimization and more.
AI in transportation
It’s impossible to talk about AI in transportation without talking about self-driving cars — it’s a both billion-dollar-plus opportunity for the automotive industry and a puzzle that is proving practically and ethically difficult to solve. Uber is using automated delivery trucks as a test case for safe self-driving; highways account for only 5% of roads in the U.S., making delivery driving a more immediately viable opportunity than consumer applications.
In addition, AI may also hold the key to better traffic management and optimized public transportation. And as one Stanford study predicted, AI is likely to have an impact on the efficiency of city infrastructure by providing better predictive behavioral models of individuals’ movements.
AI in health
It’s not hard to imagine a future where medical riddles and seemingly unsolvable illnesses are a thing of the past, and health care companies see AI as the key to unlocking answers in unfathomably massive sets of data.
For example, companies like Owlet are using smart clothing to gather hundreds of thousands of hours of infant health data a night and transport that data to the cloud. According to Owlet chief data scientist Jeff Humpherys, the next frontier is using AI to make sense of that data to bring peace of mind to parents.
“Babies are still 42 times more likely to die of SIDS while sleeping than to die in a car accident, but we still don’t have an answer for SIDS,” Humpherys said at the same SLC tech conference. “Our goal at Owlet is to figure this out. Now that we have all this data collected, we have the opportunity to know what physicians don’t. That’s something this new world of big data allows you to do.”
How to succeed in AI
If you’re just starting out, keep in mind three crucial considerations when building out an AI strategy:
According to Aryafar, establishing a quick production baseline is crucial. Often the simplest, quickest machine learning model you can deploy to production to establish correlation is the one that will yield actionable results and insights.
Additionally, some of the biggest players in AI are making it easier than ever to get started in machine learning, from no-code tools to simple APIs.
Prioritize ethical practices
Prioritizing and ensuring public safety is crucial to the sustained success of AI. While many companies are struggling with implementing AI boards, having inter-departmental discussions around organizational best practices and AI standards are still critical to using AI ethically, and making sure those who are responsible for the use of the technology are on the same page.
Get the best talent
AI roles are still taking shape as the displicine grows, so Humpherys recommends making sure your team is in a state of constant learning. Skills are perishable, and siloing your data scientists and machine learning experts can hinder your innovation before it ever begins.
In addition, having a diverse AI workforce is crucial to ensuring that algorithms don’t have built-in bias from the start.
AI is a new frontier, and it can sometimes seem safer to not engage with it all versus dealing with complicated ethical issues. But as more companies build viable, innovative use cases and expand our notions of what we can do with AI, it will become increasingly hard to ignore.