Wisconsin Alumni Research Foundation

Warf News & Media

After gBETA, UW professor’s latest startup aims to ride chatbot wave

12.18.17 | Xconomy | Jeff Buchanan | Original Publication

Virtual assistants have come a long way in the two decades since Microsoft (NASDAQ: MSFT) introduced Clippit, an animated paper clip that suggested tips and shortcuts to users as they worked in Word, Excel, and other Microsoft Office applications.

The bot, commonly known as “Clippy,” drew endless ire and mockery—even from Microsoft—before the company retired him in the 2000s. Many users felt Clippy didn’t presume competence. For example, beginning a Word document with “Dear” caused Clippy to pop up and offer help with writing a letter.

Clippy was arguably a bust for Microsoft, but newer virtual assistants have proven to be more useful and popular. From voice-controlled gadgets like Amazon’s (NASDAQ: AMZN) Echo and Alphabet’s (NASDAQ: GOOGL) Google Home to the virtual chatbots some organizations’ websites use to greet visitors, many of these tools are helping people find information and carry out tasks more quickly. And, unlike Clippy, most of these modern helpers can be easily ignored.

Madison, WI-based DataChat is one of the newest businesses attempting to ride the wave of high-tech chatbots. The company’s tools are designed to help professionals—-especially those with little or no experience writing software code—retrieve data from different sources by typing a simple query into a text box. Enterprise search is a problem that plenty of tech companies have tried to solve over the years in various ways. Other examples include Endeca, which Oracle (NYSE: ORCLacquired in 2011; Xendo, which was bought by AppDirect last year; and ThoughtSpot. DataChat’s founders think their approach could prove attractive to businesses, in part because of the software’s ease of use.

Photo of Jignesh Patel
Jignesh Patel

“The targeted user could be a decision-maker who has no programming experience,” says Jignesh Patel, one of DataChat’s co-founders and a computer science professor at the University of Wisconsin-Madison. “For example, a marketing executive might say, ‘I want to look at the latest data from Google Analytics and Facebook on my advertising campaigns. Tell me which ads are performing the best, and compare my campaigns to the performance of the campaigns from last week.’”

Patel says DataChat is designing its bots to be able to respond to requests for information that employees in specific industries are likely to have. The startup’s technology uses analytics and machine learning techniques to “decipher” what someone is asking with a particular query, he says.

“The bot effectively is writing code in real time to go and carry out the analysis” requested by a user, Patel says. DataChat then displays results in charts and graphs, he says. Eventually the startup may add an interface allowing users to ask DataChat to retrieve information using just their voices, “but in this initial phase, the product is primarily targeting textual input,” he says.

Patel previously helped launch three software companies that went on to be acquired by larger firms. Given that track record, it might come as a surprise to some that DataChat decided to participate in gBETA, a series of free, seven-week startup accelerator programs held in several Midwestern cities. gBETA’s programs are managed by Gener8tor, a Wisconsin-based organization that holds training programs for startups and invests in some of them.

“gBETA is excellent at doing two things: one is really putting you through the grinder in terms of [sharpening] your business plan,” Patel says. “The second one is [helping] us to get out in the local community as we start to engage with big enterprises in the area. Our technology can be applied across many industries. There’s so many great companies in Wisconsin and in the Midwest.”

DataChat is currently conducting “a number of pilots” with its earliest customers, says Patel. He declined to name any of them.

gBETA has graduated several dozen startups since Gener8tor launched the series in 2015. DataChat was among the 20 seed-stage businesses that completed the latest sessions last week. They were spread across cohorts held in four Wisconsin cities: Madison, Milwaukee, Oshkosh, and Beloit.

During the final weeks of the program, participating entrepreneurs pitched their startups to investors and leaders at for-equity accelerators, says Abby Taubner, one of two managing directors of gBETA.

Startups selected for gBETA do not get money from Gener8tor—unlike its core accelerator program—nor do they surrender any ownership of their businesses. Company founders get help from Gener8tor’s team and mentor network as they work to add users and customers, and ask investors to open their checkbooks.

Gener8tor recently announced it plans to hold multiple gBETA programs per year in northeastern Wisconsin and Milwaukee after receiving funding commitments from a number of large corporations, including Microsoft and Northwestern Mutual. gBETA has also been held in Detroit and the Minneapolis/St. Paul area.

Maggie Brickerman, gBETA’s other managing director, said last month that her team set a goal of positioning at least one-third of participating startups to be accepted into a for-equity accelerator or raise money from angel investors after the gBETA program. She said about half of the companies that have gone through gBETA have hit one or both of those targets.

The ultimate goal for most startups is to be acquired, something Patel knows a thing or two about. One of his previous ventures was Locomatix, a mobile data analytics company that Twitter (NYSE: TWTR) purchased in 2013 for an undisclosed amount. (The other two exits he was involved with were Quickstep Technologies, which San Francisco-based Pivotal bought in 2015, and Paradise, which was acquired by NCR in 1997.)

Patel is based in Wisconsin, but he and others on Locomatix’s team decided to have the company headquarters in Sunnyvale, CA, putting it in close proximity to Silicon Valley’s investors and technical talent.

DataChat was incorporated in May, though Patel and others at the startup began developing some of its core technology years before that. (DataChat, which now has eight full- and part-time employees, recently licensed a patent from the Wisconsin Alumni Research Foundation, he says.) Patel says that ahead of DataChat’s official launch, he faced the familiar question of where the startup should be based.

“I was given a bunch of advice to go do it in the Bay Area because it’s just much easier,” he says. “But I consciously made a decision to headquarter the company in Madison. We have the talent [here]. I really feel Madison is ripe for something big in this space right now.”

WARF