AI Startup Raises Millions to Digest Intel for Spies, Financiers and Walmart
Everything the Internet has had to say about cryptocurrency is at my immediate disposal—no Googling necessary.
Sean Gourley, the founder and CEO of Primer, a 2.5-year-old, previously media-shy startup that uses artificial intelligence algorithms to parse vast quantities of data and spit out pithy, navigable digests, is demonstrating his company’s software for me in a midtown Manhattan office near the United Nations’ headquarters. Displayed on-screen is a cascade of October headlines pertaining to Bitcoin—”Russia to regulate Bitcoin market,” “Wu-Tang Clan’s Ghostface Killah cofounded a cryptocurrency venture,” “Goldman Sachs CEO hasn’t made up his mind on Bitcoin,” and on and on.
The deluge is no match for Gourley’s synopsizing program. The tool condenses this seemingly endless list into a single page, concisely summarizing all of the important details into a general overview. The document features charts, maps, timelines, and plenty of hyperlinks for the inquisitive to drill down deeper into any particular facet of the news.
Primer’s tool spins up what looks like a well-researched Wikipedia entry on any topic: Bitcoin, terrorism, oil, organic beef, regional liquor sales—you name it. Structured and unstructured data goes in, proper analyses come out.
No more sifting through pages of Google search results to locate original information; everything is right there.
“We built a program that can read, find insights and generate human readable output,” Gourley says. “And we do that at a speed well beyond what just humans can do.”
Priming the pump
Primer is banking on businesses desiring, yet not being able to afford, their own sprawling armies of intelligence analysts to make sense of an inbound flood of digital data. So Primer offers to take the load off humans by doing the digging, the compiling, and the summarizing for them with its natural language processing tech.
Investors see potential. Primer said Tuesday it has raised $14.7 million in two previously undisclosed rounds of venture capital funding led by Data Collective, a VC firm co-managed by Zachary Bogue, husband of former Yahoo CEO Marissa Mayer. Other investors include Lux Capital, Amplify Partners, and In-Q-Tel, the investment arm of the U.S. intelligence community.
The fundraising comes just as the quantity of data generated by humans and their technologies is set to explode. By 2025, this figure is expected to ramp up to a trillion gigabytes, ten times what was created in 2016, according to researcher IDC. It’s a good time to be in business, in other words.
Get Data Sheet, Fortune’s technology newsletter
Primer’s approach has already won over U.S. spy agencies (Gourley claims he doesn’t know which, since In-Q-Tel manages the relationship with the individual agencies) and other early customers, such as Singapore’s sovereign wealth fund GIC and retail giant Walmart (WMT).
The value to clients: constantly updated briefs on geopolitical happenings (for spy agencies), market movements (for finance), or consumer products and supply chain concerns (for retailers).
“We’re operating in a fast-changing environment where there is a massive amount of data and information to understand,” Dan Bartlett, executive VP of corporate affairs at Walmart, says in a statement. He adds that the company is using Primer to “accelerate our decision making.”
George Hoyem, managing partner at In-Q-Tel, is even more bullish on the tech. “Primer’s technology is poised to revolutionize the way our Intelligence Community Partners consume and prioritize information,” he says in a statement.
If markets run on information, then Primer pitches itself as the petrol. With this tool, “the first three hours of your day are now being accelerated,” Gourley says.
From nano- to peta-
Gourley’s own academic and entrepreneurial trajectory has followed an accelerated track.
A New Zealander, Gourley studied nanotechnology at the University of Canterbury, graduating with a masters degree in 2002. Soon after, he briefly took up a post as a research scientist at NASA Ames Research Center before pursuing a PhD in computational physics at Oxford on a Rhodes scholarship.
Just as Gourley was preparing to receive his degree in 2006, he cofounded Quid, another AI startup that specialized in data visualizations. He served as the company’s chief technical officer. (Quid still exists, last raising $12.5 million a year ago from Salesforce (CRM) Ventures and other undisclosed investors at a valuation of $146.5 million, including the capital raised, according to industry tracker Pitchbook.)
A few years later in 2009, the British science journal Nature featured Gourley’s research on the mathematical structure of insurgencies in war-torn regions, like Iraq and Afghanistan, on its coveted cover. He presented a TED Talk on his findings that same year.
After a few more years at Quid, Gourley decided to set out on his own, convinced that breakthroughs in deep learning and machine intelligence warranted a new business. He began to nurture the idea that would become Primer.
Gourley’s goal is to bring the benefits of intelligence-gathering to the masses.
Few organizations have the resources to maintain big information-combing operations. You can practically name them on one hand, says Gourley, listing off the likes of the CIA, MI6, and a few big banks.
When I mention that Exxon Mobil (XOM), the world’s biggest publicly traded oil company, is known to have an impressive intelligence division (no doubt a decent training ground for U.S. Secretary of State Rex Tillerson, ex-CEO of Exxon), Gourley’s eyes light up. He notes that he sits on the board of Anadarko, a much smaller oil and gas company by comparison ($27 billion market cap versus Exxon’s $350 billion).
The glimmer in Gourley’s eye conveys that the appointment is no accident. His ambition for Primer: to sell its services to small-and-midsize companies, like hedge funds, retailers, and, yes, oil outfits. Thus the company’s Tuesday debut.
So far Primer’s team of three dozen employees—most of whom are engineers—have built the tool to handle text in three languages: English, Russian, and Chinese. Next up to be added are Arabic and Spanish, Gourley says.
“Once we build machines that help us pass our antiquated cognitive limitations, we’re going to free ourselves up to see more of the structure that exists in the world,” Gourley says.