This month, Vista Equity Partners announced the group had purchased one of the fastest growing narrative language companies, North Carolina-based Automated Insights (AI). AI specializes in developing software that takes massive amounts data and crafts it into a narrative, just as journalists and researchers do.
AI’s products are often described as “robo-journalism,” even though it’s software (not robots) responsible for crafting the stories.
It’s unclear how much money AI was purchased for. It’s also unclear what other terms of the sale are – those haven’t been publicly disclosed.
But what is clear is that the acquisition is a sign that investors are seeing a payday in companies on the cutting edge of artificial intelligence, even if just a few years ago that technology sounded like the ramblings of a crazed science fiction fan.
Vista Equity Partners specializes in acquiring tech and software companies serving different areas of the economy, including in the health, business and nonprofit sectors. The company also owns STATS, a sporting data and content company. STATS, in turn, will now hold Automated Insights, along with the previously acquired The Sports Network (TSN) and Bloomberg Sports.
As for AI, the company supplies stories and news briefs to the Associated Press. AI’s software system has been a huge success for the AP. Just last month, officials with the news agency reported that it is turning out more than 3,000 stories per quarter, a rate 10 times higher than when the AP employed only human writers.
“It’s funny,” said James Kotecki, manager of Media and Public Relations at AI. “I could say that this deal shows that natural language generation has arrived, or it’s gotten real. But it’s already arrived, and it’s been real.”
As for what this new technology might mean for the public sector, it's likely that data culling and organizing would be useful for many cash-strapped state and local governments.
Sports: The First Frontier of Narrative Language Integration?
While AI is not the only company to enter the narrative language software marketplace, the company’s integration into STATS seemingly leaves its Wordsmith software platform in a commanding position.With STATS owning TSN and Bloomberg Sports, it will allow AI a built in distribution network to publish – and promote – its unique brand of story writing.
“This deal is way more than just sports data,” Kotecki said. “Although, in some ways, [this acquisition] puts us full circle. We started by writing sports stories.”
In fact, that’s how another narrative language company, Narrative Science, began. Sports is a natural entry point for writing software because it presents large amounts of data to be summarized in a simple, linear way.
"We did college baseball," Narrative Science CTO Kristian Hammond told CNN in 2012. "And we built out a system that would take box scores and historical information, and we would write a game recap after a game. And we really liked it."
Indeed, the process of taking raw data from sports and turning it into a narrative is actually older than the modern computer. In the 1930s and '40s, sports broadcasters, including former President Ronald Reagan, would often take the box scores from baseball games and recreate them for their audiences.
Meanwhile, the French narrative language company Yseop is looking at developing more than just the straight sports story. That company has focused on making intelligence software that can actually “think” around data sets. In practical terms, it would be like if computers could take a sports score and evaluate an individual player’s performance.
Yseop's VP of Marketing Elizabeth Farabee even used a sports metaphor when explaining why that analysis is key for the narrative language sector in a recent blog post.
Stories without analysis and insight, Farabee wrote, are “like a sports play-by-play announcer without the colorful commentator.”
What Robo-Journalism Could Mean for Government
While none of the three companies have highlighted partnerships or sales to the government sector, it would seem that data culling and organizing would be a useful tool for many cash strapped agencies, especially at the state and local level.Then there’s the potential big money of contracting with the federal government. AI’s James Kotecki notes that one of the largest data sites in the U.S. belongs to Uncle Sam.
“When you look at something like data.gov, it would seem to be something we’d be looking toward,” Kotecki said. “There is so much data in the world, and to translate it into actionable narrative data would be really useful.”
Government officials could find that the software helps get projects done more quickly and at cheaper costs, much as the Associated Press has found the software to help add more content at a quicker rate.
And while Kotecki says the government sector is not a market AI has actively considered previously, that could change now with this acquisition.
“The data the government has is an embarrassment of riches,” he said.
Government officials contacted for comment on this story did not return requests for interview by press time.
