How This CIO Is Leveraging An AI Assistant To Maximize Efficiency And Savings

Schedule a call with Children's Mercy Hospital's Chief Information and Digital Officer David Chou and you will then connect with Amy Ingram, David’s assistant. After one or two back-and-forth emails, you will have a call scheduled. However, there’s just one issue with Amy Ingram the assistant.

Amy Ingram is not a person. She’s actually an artificial intelligence (AI) personal assistant.

This case is just one example of AI permeating enterprises. A 2017 study from MIT Sloan Management Review and the Boston Consulting Group found that 85% of executives believe that AI will give their businesses an inherit advantage. Meanwhile, the use of AI personal assistant chatbots in the enterprise is also growing. An Orbis Research study predicts that the AI chatbot market will grow at a CAGR of 37% by 2021.

Virtual assistants are starting to become the designated AI component for many enterprises. In the past year, Amazon (Alexa), Google (Google Assistant), Apple (Siri), and Microsoft (Cortana) have moved from homes to the workplace. Each company is vying to be the leading voice-activated AI assistant in enterprises

“Every industry is looking at artificial intelligence as probably one of the biggest buzzwords in 2018,” Chou said. “I would say a lot of organizations are using piece components of it, but true artificial intelligence takes a lot of time. You need to have a certain algorithm to perform these high level computations, and you need time to really redevelop and enhance these algorithms to really function the way you want it to.”

See Related: There is No Hype: AI Is Gaining Momentum In The Enterprise

How An AI Assistant Works

Within an enterprise, AI assistants serve many purposes. Some of them utilize speech recognition systems for email dictation or locating important documents. A few AI assistants are programmed to maximize employee productivity and improve supply chain efficiency.

Chou uses an AI personal assistant from X.AI that helps professionals schedule important meetings and phone calls. When someone requests a meeting, the user instructs the AI assistant to reach out to the person and it schedules a meeting based on the availability in the user’s calendar. If it is used correctly, the AI can assimilate flawlessly into the email conversation so that the person isn’t even aware that they are chatting with a bot.

The perfect use cases of AI assistants involve tasks that are highly repetitive and computational. As the market grows and the technology matures, AI could become more predictive and perform an increasing number of human-like functions.

See Related: Considerations For Deploying Enterprise Chatbots

The Results

Chou claims that his AI assistant successfully schedules phone calls and meetings roughly 80% of the time. The more specific the person is with instructions for the phone call or meeting, the better the response from the AI assistant. If the instructions were more broad and generic, then the AI assistant would likely get confused.

“That is the learning curve that I was referring to from people who want to use AI,” Chou said. “In the workforce, when you want to start using AI, you may have to change your behavior to be AI-ready.”

Although the there are some caveats for the time that it takes the AI assistant to learn behaviors, it will get smarter over time. Similar to voice-activated virtual assistants, such as Alexa and Siri, there are some errors that have to be corrected early on. As time progresses, the technology improves and it becomes incredibly useful.

The AI assistant augments many of the daily routine tasks that Chou’s human assistant would normally handle. Consequently, his human assistant has been elevated to a higher level role within Chou’s department. That’s a big ROI for the organization, because instead of paying a full-time salary for someone to schedule phone calls, the AI assistant automates the tasks at a fraction of the price.

Recommendations For Other Enterprises

Chou said that he first tried AI to gain another resource from a virtual system perspective, and that he has been pleased with the results. Although Chou is trying to get his colleagues to think more about AI, he is not pushing an enterprise-wide use of virtual assistance. In order to integrate AI on an enterprise-level, Chou said that there would have to be a change in behavior, and not everyone is ready for that. He would recommend the technology for other organizations that are already accustomed to AI, which Chou said might potentially be the “new normal” in the enterprise someday.

“To this day, people are using social artificial intelligence,” Chou said. “I would say overall people are thinking about using little components of AI, but we're not even close to being in that mature model. I still think it will take at least 7 to 10 years before we get AI to where it could perform almost human-like functions.”

This article was originally published on Enterprise Mobility Exchange.