Solariat’s new tool knows where the online conversation’s at

Solariat has tapped years of artificial intelligence technology to create an interesting way to market online information. The startup’s software analyzes conversations in online forums and then suggests links to related conversations.

For instance, if someone is talking about the side effects of a drug in an online forum, Solariat will automatically analyze the commentary and then suggest other conversations where the same topic is discussed. It focuses on providing answers to questions raised by people in the course of an online conversation. In some ways, it’s a big shift in the way that people can find information. You won’t necessarily have to do a Google search anymore to get the answers you’re looking for, says Jack Smith, chief executive and co-founder of New York-based Solariat, which calls itself a “conversational analysis and marketing company.” Rather, the information you are searching for could appear at the bottom of the page where you ask the question, since the web page will be updated to include links to the answers you’re seeking.

“Search queries are moving out of the proverbial box and into social media where conversations are taking place,” Smith said. “We make the conversations better.”

If Solariat does this right, then web site publishers, advertisers and consumers alike will all care about it. It is, in essence, a new kind of search engine. One of its current partners is the New York Times’ web site.

In some ways, this is as simple as searches where relevant banner ads or links are placed on the same page as your search results. But most people are ignoring banner ads these days. Solariat’s suggestions, however, are more likely to generate clicks, according to Smith. In the company’s research, users click upon Solariat’s suggestions about 10 percent of the time, compared to about 0.2 percent clickthrough rates for untargeted banner ads. Solariat’s suggestions could be ads themselves. Or they could simply be links to other conversations within the same site.

The latter is a concept known as recirculation. It generates more hits for a web site, including archived information, and thus generates more ad revenue from more page views. There are recommendation engines that do this. But recommendation engines push users in a certain direction. Solariat is more like a hybrid that both pushes and pulls, Smith said.

“We sense what people are looking for,” he said. “And we only go into the conversation when we are invited through some kind of query.”

Solariat manages to insert itself in the place where people are most engaged. That is, it only makes recommendations in the midst of conversations where someone is making a query of some kind, Smith said. Solariat supplies only the information needed, whether that is another web page or an ad.

It’s almost as if you had a human moderator snooping on an online conversation and then directing it as appropriate. But since Solariat is automated, it doesn’t invade your privacy, and it can also scale to large numbers of conversations. Solariat uses a patent-pending analysis technology that analyzes a conversation. Then it automatically generates a message to point the way to the answer the question at hand. Solariat’s adLib technology automatically designs ad creatives that serve as helpful, attractively designed pointers to the answers. It surfaces content that isn’t viewed as much as it could be, and it increases an advertiser’s ability to participate in a conversation with the consumers it is pursuing.

Smith says these answers mesh with the conversation. To bring up another example, if a consumer is participating in a discussion and posts the question, “I love those Reebox EasyTone shoes. Anyone know where I can see all of the different styles?” Solariat would interpret that query, build a response such as “find all of the styles of Reebok EasyTone shoes on this page,” and include a link to an information page from Reebok.

The technology is based on artificial intelligence research. Jeffrey Davitz, chief technology officer at Solariat, learned all about the rules of conversational media as the creator of the social computing group at SRI’s artificial intelligence center. He built a system for the U.S. Army that made conversations better and was on the management team of CALO (Cognitive Assistant that Learns and Organizes), the largest artificial intelligence project in history that involved dozens of companies and government research groups. Solariat does not license anything from SRI or CALO. But one similar spinoff from CALO and SRI was Siri, a voice-driven app that serves as a personal assistant to make restaurant reservations. Siri debuted in February and was purchased in April by Apple.

The artificial intelligence is useful with social media recommendations because people don’t like to have their conversations disturbed for no reason. If they get irrelevant messages, they get annoyed. But if you serve them the right information, they’re happy. Google discovered this with the ads on its search results. But traditional ad links don’t work that well in the midst of a social media conversation. Solariat is thus following users as they migrate from finding answers on search pages to finding answers in social media.

Smith is a veteran of internet companies WPP and 24/7 Real Media. The company has raised $1 million in funding from KPG Ventures. The company was founded in November, 2009 and has seven employees. Rivals range from human-powered search engines such as Mahalo to Google’s own search engine.

VentureBeat is studying mobile marketing automation. Chime in, and we’ll share the data.