davidl
David Lee
October 14, 2016

AI: The Dream and the Reality

AI was the big theme of this year’s Dreamforce, where I spent most of last week as both a presenter and an attendee. No other annual tech gathering matches Dreamforce in excitement and visionary panache, and for good reason. Salesforce CEO, Marc Benioff and his team have always been prescient when it comes to identifying—often years ahead of time—the next big thing. They presaged the importance of enterprise messaging, team collaboration tools, and IoT, among other trends. When they predict that something will shape the technological landscape of tomorrow, it usually does.

As someone working in the unified communications space, I share Benioff’s belief that chatbots—many, but not all of which incorporate AI— have a lot of interesting potential. We also share the perspective that AI’s main value lies in helping humans do the things that humans are good at, while turning those other tasks over to machines. The promise of AI is that it will take a large set of data, quickly extract predictive insights, and then use those insights to automate those business processes ripe for automation.

The end goal is higher efficiency, and we’ve followed this approach at RingCentral, driving efficiency through iterative improvements—many involving automation— to our platform. That goal was also behind our acquisition and integration of Glip: We realized just how much of business communication today takes place in the messaging space, and that enabling businesses to pass information through Glip, where it is easily consumable, would have a powerful efficiency impact. Today, we’re working to add even more business value to Glip workstreams by developing chatbots that can help users with a variety of tasks.

That said, the field of AI still has a long way to go before many of the capabilities envisioned become available in a meaningful way. As someone who leads product strategy and drives the roadmap for our developer platform, I’m keenly aware of the work that still lies before us if we are to make AI-enabled chatbots, and other AI-driven technologies truly useful in a business context. Right now, there seems to be a lot more dream than reality.

There was an interesting moment at Dreampitch, Dreamforce’s Shark Tank-style startup pitch competition, that highlighted this lag between what AI is today and what it will be in the future. Claire, the winning startup, uses chatbots delivered through Facebook Messenger to engage with target customers and evaluate how they respond to various consumer products.

When the judges asked Claire’s cofounders why their bots don’t use natural language processing, or NLP—a distinctly AI method of communicating and something of a buzzword—they responded that NLP is “a little too early to use” in their product tests, and were emphatic that theirs is not currently an AI-based service and instead relies more on a pre-scripted experience. I believe this is an example of an innovator taking a realistic and measured approach toward these emerging technologies where the rubber hits the road.

While we don’t quite know when it will be possible, what will be possible in the AI-enabled unified communications space down the road is very intriguing. I believe that one area that holds a lot of promise is sentiment analysis on voice. Because voice is the mode of communication most heavily used on the RingCentral cloud platform, we have a wealth of customer voice data at our fingertips. In a sales environment, for example, AI capabilities could help index all of the conversations in a company’s CRM solution and pick out the keywords or phrases that best helped close the deal. Each word could be assigned a sentiment score, and sales reps could inject the highest-rated words into their dealings with customers in order to have more efficient conversations—in real-time and without any input from other reps. Insights gleaned from this data could be incorporated into training and best practices, further amplifying its impact.

It’s possibilities like these that keep the AI dream going strong. Without a doubt, when AI does arrive to our workday realities, we’ll be more than ready to welcome it.