Andi Gillentine is the co-founder and COO of Austin-based Whit.Li, the site analyzes words people post on social media sites like Facebook to determine their personality traits.
Whit.Li is the only Austin company selected to pitch at the highly selective Women 2.0 conference. Gillentine will compete Wednesday against nine other women-run startups from across the country in a day-long competition. The other startups include ActivityHero, Alike, Citizen Made, Clear Returns, EatDrinkJobs, LightningBuy, Maternova, NewlyWish and The Daily Muse.
The winning startup at Pitch NYC 2012 will receive a meeting with Marc Andreessen, co-founder and general partner of Andreessen Horowitz, a one-year hosting service to Rackspace, an automatic finalist interview for TechStars NYC Winter 2013 class and a L’Oreal Women in Digital grant of $25,000 in cash.
Whit.Li has specialized natural language processing technology that gives businesses a better understanding of who their fans and followers are online.
“It took nearly a year to get the core technology built and tested. We initially focused on Facebook data as a source. To gather our first data to test our algorithms, we used a Facebook app where we provided a personality analysis based on a user answering a Big 5 personality inventory and authenticating that we could look at their posts,” Gillentine wrote in this post on the Women 2.0 site. “We used that data, anonymously, to refine our algorithms to provide personality, demographic and interest analysis. In keeping with our original inspiration, we released the technology via API so that companies could build it into their sites directly. We launched our API at SXSW 2012. Initially, we thought that collaborative consumption companies would be very heavy users of this technology. At, SXSW we learned that brands – big, enterprise brands – were even more interested in it. That was an eye-opening experience for us.”
Whit.Li now has 200 companies using its technology and another 200 waiting for the next version, according to Gillentine.