Smarter.Codes is in its 3rd generation now and things are looking exciting now. Here is a pilot run we did where we were provided with some less than 100 documents shared among customer support department of a company (essentially their knowledgebase compiled over years of operation).
This company has 4,000 clients, actively calling their customer support call center that has 60+ customer support executives. Finding appropriate answers out of their knowledgebase at times becomes a challenge (given that those are a lot of documents since the company has 400+ products). The customer support was slowing down and support agents spent alot of time looking for the right person to answer a question rather than looking into 100 something documents.
Smarter.Codes AI digested all documents into structured data and was turned into a chatbot using Skype network (although, it can be made available on any / all networks like Slack, Gtalk, IRC or even email or sms).
What is special about this bot?
1. It can read normal knowledge documents, PDFs, CSV files, Excel sheets, presentations or even emails and past chat conversations and start answering questions about them right away.
2. Not just documents, it can look into large / Big database and perform queries to support its answers.
3. The bot can learn overtime (using machine learning). The question it is not able to answer is raised as a flag in the backend that it can be trained upon.
4. Furthermore, any answers that are most commonly drilled down in a particular session become the default answers since they are the most commonly looked for answers by customer support / users.