If you watch TV, you’ve probably seen the Microsoft commercials that take potshots at Siri. Ditto from Google, pushing Google Now. But it’s not just the big names of tech who are running in the digital assistant race: Plenty of smaller companies are also competing for a piece of that sweet predictive analytics pie, and one of them looks especially formidable.
Expect Labs is a small, San Francisco-based startup whose flagship product is an API called MindMeld. A combination of voice recognition and machine learning technology, MindMeld is a big bet on the Internet being navigated through automobiles and cable boxes rather than touch screens or desktops. Expect Labs sends crawlers through clients’ websites which build custom knowledge graphs, provide voice recognition client libraries for developers, and then use MindMeld’s API for natural language processing of user queries. In not so many words, it adds Siri-like capabilities to any website or app.
The API launched in early July; cable provider Liberty Global is on record as one of their early clients. CEO Tim Tuttle told Co.Labs that some of the company’s other backers include Google Ventures, Samsung, Intel, and Spanish communications giants Telefonica.
Shortly after we spoke, Expect Labs received funding from an unexpected source: American intelligence agencies. In-Q-Tel, an investment arm of the CIA and the United States’ other three-letter security agencies, entered into a strategic investment and software development agreement with the company. The terms of the In-Q-Tel investment are not known but according to the press release Expect will help "voice-enable a wide range of potential applications for use across U.S. Government agencies."
Prior to the In-Q-Tel funding round, Expect Labs is believed to have raised approximately $2.4 million. "Voice is a very useful input for new sorts of devices like set-top boxes or Apple TV," Tuttle added. "We’re also looking at a lot of connected car applications. There’s a huge investment in connected voice applications inside cars, and voice-driven intelligence systems for cars that accomplish tasks while driving." Other areas of interest, he added, include helping customers navigate large video libraries or product catalogs through tablets for their clients. Expect Labs feels that adding voice queries to video-on-demand and watch-through-tablet services makes finding content easier than wading through multiple menus.
Other companies are also looking to add personal assistant components to third-party products. Another Bay Area company, Speaktoit, is funded by Intel and Motorola and is positioning their API as a tool for automakers and in-car app makers. According to a Speaktoit fact sheet, the company can build "vehicle-specific dialog interfaces," and already has a dozen OEM partners in the automotive world. It’s important to note that both Expect Labs and Speaktoit both offer their product in multiple languages; Spanish, Korean, and German are just as relevant to the personal assistant world as English is.