TalkTo raises $3 million for app that lets you text questions to businesses

Getting fast answers to users' questions is a complex proposition. TalkTo will endeavor to make that happen by relying on a hybrid of software and human intervention. The company employs full-time "operators," who work at home, to help handle questions and make the software smarter. (Most of them have several years of experience in call centers, Levinson says.) The first time you try to ask James Hook & Co. about today's per-pound price for lobster, for instance, an operator might get in touch with the business to get the answer, and also find out the best way to forward future questions. Does the business prefer e-mail, live chat, or perhaps a text message to the owner's mobile phone? Then, future questions will be sent that way, with operators only intervening if there isn't a response. The company also plans to do a good deal of data collection from websites (a/k/a "scraping") to be able to answer common questions — like when the Starbucks in your neighborhood closes, or whether Best Buy carries Bose brand headphones.

Levinson says TalkTo is "part human, part machine." The human part sounds expensive to me, especially if the app really takes off. But Levinson wouldn't get into detail about how TalkTo plans to make money, saying only that "we're not going to be religious about any particular monetization method today." He and Crane did make two interesting comments, though. First, they observed that people who are asking specific questions about products in stock and the soup of the day tend to have a pretty good intention of showing up to spend money. Second, they said that just as well-reviewed businesses on Yelp tend to attract more customers, so will businesses that respond quickly to questions from prospective customers on TalkTo. Crane says, "If I can see that one locksmith typically responds in ten minutes, but another takes six hours, I'm probably going to choose the first one if I'm locked out of my house." Sounds like TalkTo could either provide tools to businesses that make it easy for employees on duty to respond to questions, or highlight in the app businesses with zippy response times. The company might also charge businesses to send outgoing messages to consumers who have texted them in the past: "You asked about Sony TVs last month, and we just wanted to let you know that they're now on sale," for instance. Or perhaps give competitors the opportunity to acquire new customers: "Rami's Falafel may be closed right now, but Falafel Palace is open 'till 3 AM."

For the initial launch phase, the app will only be able to be activated by people in the Boston area (though you'll be able to contact businesses anywhere.) Levinson says he wants to be sure the app serves users in one metro area well before expanding to others later this year.

Before starting TalkTo, Levinson was an entrepreneur-in-residence at General Catalyst, the Cambridge venture capital firm. In 2004, he sold a content management company to IBM. Crane is on leave from a fellowship at MIT's Media Lab, where he led the team that won DARPA's red balloon challenge in 2010. (Crane is pictured at left, with Antonio Rodriguez of Matrix Partners)

The company has raised $3 million in funding so far from Matrix Partners. Antonio Rodriguez of the Waltham firm says, "I came to Matrix to focus on consumer investments, but it has taken me two years to find a consumer play that I'm excited about." (He previously put money into B2B startups The Echo Nest, which provides technology to developers of music services, and Adelphic, which is devising technology to deliver mobile ads more intelligently.) The funding, Rodriguez says, ought to be enough to enable TalkTo to "figure out how to get to scale and monetize" its customer base.

The eight-person company has been operating so far out of Matrix's Kendall Square outpost, but is currently hunting for new office space.