A unique application of deep learning in handling customer complaints and questions

Almost every site requests customers to submit their requests via a form. Most businesses can't afford a call center. Even the ones that do obscure their numbers because they want to minimize the number of calls and hence human operators. Which means that contact us forms are here to stay. However consider the default experience of contact us forms. The user sends a message "I want to know if you do a free pilot of your service?" Or maybe they write "how can I get access to my tax documents for 2015?".

The standard email response is that "we got your message and we are looking into this". But can we do better?

AI especially deep learning has been shown to have some success in automated email replies. Imagine applying the same principle to contact us or report an issue form. To train such a system you can simply feed in all the previous responses to all the previous questions asked. The deep learning system can then for new requests auto generate an email with the response as written by a computer.

So a sample response might contain a link to a relevant article, or might contain a follow up question.

It's important that in addition to automated response there should be a link saying I want the talk to a operator my issue is stilll not resolved. The benefit of this system is that it can significantly reduce the number of emails that need hand written response. All your previous responses will continue to grow big and the model should be updated on a regular basiss with new questions.

If you would like to test drive a solution like this please email at info@sublimeai.com


Popular posts from this blog

Large SAP tables which are good candidates for partitioning

Partitioning in SAP HANA database