Companies who have been running a bot for a while will realize that the main reason why bot doesn’t understand users is because they write so badly, they make typos and mix languages. Being honest, users don’t make it easy to be understood. This is the challenge that bot’s developers have to face.
It’s 10pm on Saturday. I am having troubles renewing my car insurance online and I need customer support right now. There will be anybody attending me? Or are you willing to lose me as a customer?
Your hotel deals with thousands of reviews: in TripAdvisor, Yelp and your own satisfaction surveys. Nobody can process all this information and get accurate insights of what is going on. So, companies start relying on machines to do this job automatically by using sentiment analysis.
Bot accuracy can now be increased, easily. How? Increasing training data, not by hand, but using automatically-generated query variations. We have benchmarked Rasa and other platforms and accuracy goes up to 93% with Bitext artificial training data tech.
A client came to us because their bot couldn’t understand customer requests like: “I’ve been a client for five years and my daughter wanted that present so badly; but I haven’t received it yet!". It was impossible for their bot to understand such a combination of knowledge items. That’s how we started implementing query segmentation.