Waverly

Philippe Beaudoin knows what you’re thinking – it’s part of his job. As a technologist with experience in AI, Philippe had previously worked for one of the biggest search engines in the world. He understands intimately how, when, and why we go to find the information that we’re looking for, and how technology is used to curate those search results and feed us the answers to our questions.

Unfortunately, far too often that recommendation technology can become maligned and misused. When that search engine has accumulated AI where it can ‘guess’ what the user is looking for, it can lead down a road to dangerous results, such as by creating something of an ‘echo chamber’ through filter bubbles, or by leading users towards increasingly radical ideas. Both have a strong potential for inaccurate information, and neither truly benefits the user.

Philippe realized that there was an alternative, and has since begun building out Waverly. “My experience in natural language understanding (NLU) convinced me we could build recommenders that would be based not on opaque user behavior, but on the understanding of user’s intentions — expressed through natural language, in their own words,” said Philippe. Waverly categorizes information through ‘waves’ – short text that users create on a particular topic, and can then collate articles that might be a good fit pertaining to that topic.

While still in the development stage, Philippe says that the company is on track to launch its mobile reader application later this year, and then leverage user data to keep improving the algorithms. He’s proud to be working with K-Man throughout this journey. “As we grow, I’m convinced K-Man will be able to help with many aspects of scaling a startup,” said Philippe. Waverly is making plans to become a universal recommender system that can be incorporated into any platform, and K-Man is excited to be along for the journey.