Answer engine JustAnswer now uses bots to route questions to human experts
The use of bots to qualify questions or leads, which are then handed off to experts or reps, may be the optimal chatbot use case.
JustAnswer is one of the survivors of the “answer engine” or Q&A craze that was prevalent a number of years ago. The venerable (or ancient) Yahoo Answers is still around, and so is Quora, but various efforts from Google, Facebook, Amazon and a range of startups are gone.
The most recent entrant, Biz Stone’s Jelly, was acquired by Pinterest earlier this year.
The pitch is compelling: Humans are better than algorithms at answering complex questions, and users want “answers not links.” Yet almost nobody has been able to get the formula right (quality + scale + a business model) — and that includes Yahoo and Quora. But JustAnswer has managed to make it work.
Founded in 2003, JustAnswer adopted the paid-advice model that was also used by the original Google Answers. Each user who connects with one of 12,000 experts on the site pays on average $30 for a consultation. There are no ads. Most of JustAnswer’s traffic comes from SEO.
Last week, the site introduced “Pearl,” a virtual assistant intended to answer simple questions and qualify leads for the site’s roster of experts. The tool has been in beta testing for three years and has the advantage of being trained on 16 million questions and answers in the company’s database.
“This is a killer app for the chatbot era,” says Andy Kurtzig, JustAnswer’s founder and CEO. Kurtzig says Pearl adds significant efficiency to the process, eliminating the need for the experts to spend as much time determining the nature of a consumer’s problem or question before responding. Pearl can operate as an intelligent routing engine.
The bot can recognize more than 100,000 variables in conversation and ask context-specific follow-up questions. For example, if there’s a pet problem, the assistant would seek to diagnose the problem generally and then forward that information to the veterinarian-expert before the consultation. Users are asked during the conversation if they want to talk to an expert.
This not only helps isolate the question or problem quickly, it likely improves close rates and the percentage of consumers agreeing to pay for advice. “This is the future of professional services,” says Kurtzig.
Several years ago, Google introduced an expert video-chat platform, Google Helpouts. It was very promising, and one might have assumed Google’s visibility and resources would make it work. However, in 2015, after disappointing adoption, the company shuttered Helpouts. Another take on expert content, Google Knol, was also shut down after a few years.
Despite these setbacks, companies continue to try to deliver customized advice online. Bots and machine learning may make that more feasible than in the past.
Last week, I wrote about how Valassis used Facebook ads, Messenger and human sales reps to drive offline car sales and leases. This hybrid use of bots and humans, similar to the JustAnswer approach, is the conceptual model for optimal bot deployment, I believe.