Interview with Robbie Allen, CEO, Automated Insights
For podcast release Monday, July 28, 2014
KENNEALLY: In science fiction, robots make lighting-fast calculations and have superhuman strength. From Forbidden Planet to 2001: A Space Odyssey, the robots of cinema have played a wide range of roles as friends and as enemies to humankind. But of all the wonders robots have performed, until now, no one has imagined them as journalists.
Welcome to Copyright Clearance Center’s podcast series, I’m Christopher Kenneally for Beyond the Book. When AP announced recently that data driven stories about company earnings would soon be written automatically, AP managing editor Lou Ferrara stressed the move was not intended to replace living, breathing writers with heartless machines, but to free up his staff to do more reporting, going beyond the numbers to provide insights on what the numbers mean. The remarkable technology that produces personalized narrative content from big data is developed by Automated Insights, and the company’s CEO and founder, Robbie Allen joins me now from his Durham office. Welcome to Beyond the Book, Robbie.
ALLEN: Hi, Chris, thanks for having me.
KENNEALLY: Well, we’re delighted you could join us today. We’ll tell people briefly about your background. A veteran of Cisco and IBM, Robbie Allen is the author of 10 books on a variety of technical topics from O’Reilly, Apress, and Addison Wesley. His company, Automated Insights, helps other companies in such markets as finance, fitness, business intelligence, real estate, and sports, to realize the full potential of their data. AI’s technology is able to humanize big data by personalizing stories to one user among millions delivering real-time, tailored narratives.
Robbie Allen, people listening to us today might wonder if they’ve ever read a story written by a robot, and the chances are increasingly that they have. In fact, earlier this year, when a light earthquake hit the Los Angeles area, the very first reporting the Los Angeles Times published was the same type of automated journalism. So tell us more about what’s behind all of this and just how common it’s becoming.
ALLEN: Sure, and Chris, I bet you most of your readers or listeners have read automated content and probably didn’t even know it. The quality of the content has gotten significantly better over the last couple of years to the point where Yahoo recently did a poll, essentially asking users to compare or determine which content out of human-generated versus machine-generated was done by human or machine. Effectively, the results were people couldn’t tell the difference.
KENNEALLY: And I bet a lot of work goes into making that possible and making differences as small as possible. You began this kind of remarkable work by generating sports stories because I think you’re a sports fan.
ALLEN: That’s right. I thought for certain types of quantitative analysis that you could actually automate that. You could create algorithms that perform similar types of analysis to what people did, and then describe what was interesting about that analysis in plain English.
KENNEALLY: Indeed, and so far your company, Automated Insights, is already well into writing something like a billion stories this year, which is an increase of almost threefold from last year. So what’s making it all popular? Is it the amount of big data that’s now available that’s making this possible? Or is it just kind of a thirst for content.
ALLEN: Well, to get those type of numbers, it really comes down to adding personalization because what we can do that just hasn’t been possible in the past is actually create a piece of content that’s tailored to an individual. The traditional media model is expert-driven, that is, you’re going to create a piece of content and hope that as many people as possible will read that or listen to it. Our model is essentially the inverse of that. We create a piece of content that we intend to have only one person read, but we can do a million versions of that story that’s tailored to each individual user.
KENNEALLY: So if I understand right, these stories are written on the fly, as it were, based upon searches. Is that how it works?
ALLEN: Yeah, it’s a combination of using cloud-based architecture and real-time data sets that we collect either from third parties or directly from customers, and generate millions of stories in a matter of seconds.
KENNEALLY: And when I get such a story, when it appears on my tablet or my PC, will I know that it was written on the fly by a computer?
ALLEN: Oftentimes, there’s not any direct indication. If you’re reading and you sit down and think about how is it that a large company was able to produce something totally tailored to me in the timeframe they were able to do it, you could probably figure it out and in some cases, companies put a tagline about us, where it’s powered by Automated Insights, some don’t. So really you never even quite know if you’re reading content that was generated by us or not.
KENNEALLY: And the kind of content, what caught our imagination at Beyond the Book, at least, was that AP had gotten into this business. But this is content today that’s being created by all sorts of companies for marketing purposes and not simply for journalism.
ALLEN: Absolutely. We sort of span the spectrum now from media companies to business intelligence solutions, to sales reporting to marketing analytics – really anywhere where there’s a significant amount of data and people want to tell stories about what’s going on with that data, there’s an application for our technology.
KENNEALLY: Your technology – you call it the Wordsmith platform, and what we are speaking about here isn’t only text, if I understand right. You can also help create visualizations of that data.
ALLEN: Absolutely, especially in the business intelligence space, visualizations have long been the primary means to communicate insights about data. We believe narrative and using text is typically a better way to do that, but we also believe in the power of visualizations, so oftentimes our solutions combine the two. We’ll have a narrative that describes what’s interesting about the data, and then use visualization to supplement the story as needed.
KENNEALLY: What’s interesting about this for me is, again, the partner of journalists these days is marketing and public relations. So agencies that would have wanted to get the word out about their companies into as many (break in audio) as possible to spread the word, if you will, must find this kind of technology very attractive.
ALLEN: Yeah, again, we’re able to help companies reach an audience that they just weren’t able to in the past. Typically, most companies associate the volume of content that they can produce with how much money it takes to produce it. What our technology enables on, specifically, the Wordsmith platform is we can remove those barriers to the extent that now you can reach, at least in our minds, to generate content for 100 users is about the same amount of energy as to generate content for a million users, and so we encourage our customers to generate as much content as they can see a use case for.
KENNEALLY: We’re chatting today with Robbie Allen, the CEO and founder of Automated Insights, and talking about journalism. We say it’s written by robots, but really is written, as you say, by the algorithms, and you’ve been working on this for a number of years. So the insight for you was the sports reporting that you wished you had seen But when you began working on this Robbie, did you imagine you would be working with this respected organization, the Associated Press?
ALLEN: You know, I’d hoped that eventually we would obtain that. I didn’t think it would happen quite as quickly as it has. In fact, that’s the thing I’ve been most surprised about is how quickly the technology has evolved and improved to the point where you have an establishment like the Associated Press that’s willing to embrace it the way that they have.
KENNEALLY: And when they have embraced it, and I mentioned in my into they have pointed out that this is not about replacing those living, breathing writers with machines, but to free up the staff to do the kind of reporting that really only a human being can.
ALLEN: That’s right, and I think we’ve never replaced a job with Wordsmith. All that we’ve ever done is essentially augment what news organizations, in this case, have been able to cover. Again, our main focus is helping organizations drive personalization down to the user level, but then also broaden coverage like we’re doing with the Associated Press, where they’re going from covering 300 companies a quarter to over 4400 a quarter.
KENNEALLY: I’m sure that rather like Google, with its translation algorithms, they are able to improve as they go along, and so it gets better and better with time. Is that what’s happening with your solution?
ALLEN: That’s absolutely the case. As we’re able to access more historical data and as we’re able to capture the type of content that’s created in every iteration, we’re able to keep track of that and automatically learn and help the content get better as it goes on.
KENNEALLY: We’ve been chatting today with Robbie Allen, he is the CEO and founder of Automated Insights in Durham, North Carolina, and we appreciate your joining us today on Beyond the Book, Robbie.
ALLEN: Thanks, Chris.
KENNEALLY: Beyond the Book is produced by Copyright Clearance Center, a global rights broker for the world’s most sought-after materials including millions of books and e-books, journals, newspapers, magazines and blogs, as well as images, movies and television shows. You can follow us on Twitter, find Beyond the Book on Facebook, and subscribe to the free podcast series on iTunes or at our Website, beyondthebook.com. Our engineer and co-producer is Jeremy Brieske of Burst Marketing. My name is Christopher Kenneally. For all of us at Copyright Clearance Center, thanks for listening to Beyond the Book.