We're building a natural language processing bot capable of working with eCommerce stores inventory and product ranges / attributes. Currently learning with Amazon lex - but always looking for more research and experts in the space.
On my AI course, I usually give the following advice to managers who wants to build customer-facing chatbots: Don't do that unless:
Although the NLP techniques are progressed significantly - recently, thanks to GPT and Bert - they still very quickly lose context in real-world conversations, especially in the eCommerce context. So the user experience will bad, which is a risk you might not want to take. Based on my experience, chatbots, however, can work very well as an internal tool that helps the customer rep.
Hey - thanks for the reply.
How would you narrow the context? Is the context not already narrowed by the fact you’re speaking to X brand in X industry? For example a coffee brand - the queries and questions are (likely) already extremely limited compared to a general chatbot.