Bridging the Communication Gap Between Human and Machine

Just as the ease of using voice has led to the growing development of smart speakers, so the use of language will lead to ease of data base analysis and interpretation. JL

Erin Winick reports in MIT Technology Review:

There has been an idea that because there is a lot of data sitting in organizations, by giving users access to that data, they will acquire the skills necessary to analyze, interpret, and act upon it. It would be easier to teach machines to communicate in our language than it would be to teach everyone in the world to interact with computers and take advantage of all the data that’s available. AI will be ubiquitous. There will be less time spent cranking through spreadsheets and pouring through databases and more querying data and getting answers back. More conversational experience with data
Artificial intelligence is seeping into an increasing number of industries, like finance and manufacturing. Now Chicago-based Narrative Science is successfully bringing AI into writing. Founded in 2010 to automatically turn statistics into baseball stories, the organization has evolved into a powerhouse in natural-language generation.
Stuart Frankel is the CEO of Narrative Science and has helped guide this transition from sports statistics to business insights. We spoke to Frankel about how technology like this is changing daily workflow in different industries and bridging the language gap between human workers and machines.
This article is part of a series of Q and As paired with our newsletter Clocking In, which covers the impact of emerging technology on the future of work. Sign up here—it’s free!
Narrative Science got its start turning statistics into news stories. How did this help you build your business and train your software?
That was really the impetus for getting the company going. We licensed the technology in 2010. We started to write baseball stories. We were able to do baseball stories, financial news stories, and real estate market roundups. We really started to build a business in media, but over time we shifted toward being an enterprise software company.

Stuart Frankel, CEO of Narrative Science.

Narrative Science

What prompted this change in focus?
We started to get a lot of interest from people who heard about us from our work in media. I always joke that if you want to get a lot of press as an early-stage company, do something that is perceived to disrupt journalism, because reporters love to write about their own industry. It helped build awareness of Narrative Science to the point where we had lots of inbound inquiries across many different industries describing what was essentially the same problem. These organizations were sitting on a lot of data.
To fast-forward, we now have about 100 customers. The work we do for these organizations and their uses falls into three broad buckets: operational efficiency, increasing customer engagement, and compliance.
Are your new customers mainly coming from one specific industry? Within the last few years, about 60 percent of our business is in financial services. So we work with companies like USAA and MasterCard and Franklin Templeton and a number of other large financial services organizations.
How are these companies able to take advantage of natural-language generation?
There has been an idea in the last several years that because there is a lot of data sitting in all organizations, by giving users access to that data, they will acquire the skills necessary to analyze, interpret, and act upon it. We’ve always felt that was ludicrous, to expect everyone in the world to acquire the skills of a business analyst or a data scientist. We felt it would be easier to teach the machines to communicate with us in our language than it would be to teach everyone in the world to interact with computers and take advantage of all the data that’s available.
What are the industries that should be taking advantage of artificial intelligence that have yet to fully embrace it?
I think ultimately, AI is going to wind up being ubiquitous and impacting every industry. Whether it’s finance or retail or health care, there is an enormous amount of data now. There are various constituencies who need information that can be gleaned from that data, either for just information purposes or to make data-driven decisions.
How do you envision work changing as a result of the new AI tools?
There will be less amount of time spent cranking through spreadsheets and pouring through large databases of numbers, and more querying, for example, large data sets and getting answers back. More of a conversational experience with data as opposed to the way it works today.