Which TV Hit Will Go Viral Next? This Israeli AI Startup Already Knows

The call for more diversity in television and movies has been a much-talked-about topic among under-represented groups worldwide who are not satisfied with the current levels of diverse representation on screen. Meanwhile, an Israeli-founded startup based in LA has seen an uptick in diversity and inclusivity as well as more consumption of foreign content than ever before. And they have the data to prove it.

Vault AI, a company using machine learning to gather insights from the audience and predict which content will engage viewers and do well on screen, has recently noticed that content featuring diversity is the content that consumers are looking for.

“I think on a global level there has been more lip service given to the diversity movement,” David Stiff, Vault AI co-founder and CEO, tells NoCamels, “Social justice has been shining a light on creating more diverse and inclusive stories. We’ve seen more Black content, more Asian content, more content featuring people of different backgrounds coming through our platform. One of the core things we’ve seen is that it doesn’t matter where the content comes from anymore. If you’re American, it doesn’t mean you want American content. If it’s good content and it’s coming from Korea, people will watch it.

Squid Game anyone? “I couldn’t imagine a show [with subtitles] like Squid Game being the number one tv show in Australia,” explains Stiff, an Australian native. “But now it is and it’s amazing.”

If you’ve ever been to the movies or streamed a television show, chances are you’ve seen Vault AI at work. Stiff tells NoCamels the company built a dataset that has over 50,000 movies and tv shows from 60 countries and developed a proprietary way to analyze the content. The content is analyzed to reflect feedback through more than 150,000 data points linked to story, talent, production, and more. It then compares it with the company’s catalog.

Clients that include Hollywood studio and marketing execs have used to Vault AI to test thousands of pieces of content and help with pilot testing, concept testing, and testing of trailers.

“So if you’re thinking about what’s going on in this interview right, it’s one person that’s in Israel, one person in the US, one’s American, one’s Australian. One’s a journalist, the other is a CEO of a tech company. These are all the different things that the platform picks up from all the different content we have. We’re looking for 100,000 different story elements that exist within any given title right? So we have 50,000 titles as well as a huge amount of data from ratings in the US to demographics in the UK, Australia — we have lots of performance data.”

How it started

In 2015, AI was just a cool buzzword, Stiff tells NoCamels. The Australian native immigrated to Israel with a past that involved work in film, television, and technology which he eventually parlayed into work for Israel’s local high-tech industry. He fell in love with the sister of one of Vault’s other co-founders, Nir Tzahar, and they decided to start a company with Tzahar’s former co-worker Ziv Ayalon. Both Ayalon and Tzahar had expertise in machine learning. Today, Ayalon is the company’s CTO and Tzahar is the chief scientist. The company employs about 30 people, with the entire R&D department in the Israeli coastal city of Netanya.

The company began by cold calling and emailing Hollywood studio executives to tell them AI could help predict the success of their next movie.

“You could basically email all these different studio heads and they’d respond to you if you had a cool story. Somehow, we got meetings with a lot of people. My first trip to LA was a meeting with 10 really, really high-level executives at the studios. We’d pitch to everyone and somehow we didn’t get laughed out of the room. But there was definitely a lot of skepticism around what we were doing,” Stiff explains, “And they threw a bunch of scripts at us and they’re like, ah, kid, you’re up. Let’s see what you got. So we started out doing a lot of free insights and that kind of thing as a beta program. And yeah, we monetized pretty quickly. It’s been going well ever since.”

Stiff says the company predicts thousands of different titles every year and can help all aspects of tv and movie production. “A lot of the content comes through our platform and immediately helps executives on the marketing side understand who they should market their tv and movies to. And then on the very early production side, it helps production executives understand what changes they should make and who the core audiences are. So when they’re making their titles, they can keep this in mind at a very early stage.”

The company also touts its accuracy and accelerated turnaround times compared with longtime research tools like surveys or focus groups.

Vault AI closed an $8 million Series A funding round in December, co-led by Israel’s PICO Venture Partners and Hearst Ventures, the investment arm of Hearst Corp., and an early investor in companies like Roku, Pandora, and Buzzfeed. Other participants in the Series A included former Sony Pictures Entertainment Vice Chairman Yair Landau, as well as existing Vault investors TV Azteca and Sky/ProSieben-backed REimagine Ventures.

“Consumer habits formed during the Covid-19 pandemic are supercharging shifts in the media market on multiple fronts,” said Gil Canaani, managing director of Hearst Ventures and the newest member of Vault AI’s Board of Directors said. “Competition for subscribers and viewers is fiercer than ever and as a result, the need for AI-powered, real-time and accurate consumer insights data has soared. Vault AI’s team combines the best minds in AI from Israel with world-class consumer insights talent, uniquely positioning them to fill this need and empower media firms to make smarter content decisions faster.”

Subscribe to NoCamels weekly newsletter and get our top stories