Automate NLP Quote Requests
The courier, parcel, and express mail division of a large logistics company was looking for a partner to help improve their customer experience around quotation requests with the help of intelligent data recognition.
Currently, these requests are mostly manually processed given that they reach the company through a variety of channels such as email, personal notes, phone notes and fax. Not only does communication come through different channels, but the structure and information level of these requests also varies widely.
A lot of time and effort is spent on these types of customer requests with low productivity, resulting in long waiting cycles for customers up to the point that sometimes the request is never even answered. Not only is the customer experience negative, but potential business opportunities are missed. It was determined that intelligent data recognition (IDR) technology should be utilized to process unstructured content from email requests.
The use case selected as a first proof of concept was around a customer quote request that could be automatically answered with the correct pricing for the shipment of a variety of packages (number and weight) from different origins to different destinations on a particular date.
Our experts leveraged the power of artificial intelligence (AI) in Hero_Flow and created two AI models to solve the problem. The first analyzed incoming emails to find out the intent and automatically selected the ones that were identified as quote requests.
After filtering out the quote requests, the second model analyzed the unstructured text of each email to find out the quantity and weight of the packages the user wanted to ship, the origin and destination and the requested date of shipment. Once this information had been extracted, Hero_Flow looked up the price in another system and then composed and sent a reply email to the customer to inform them about the availability and cost of their planned transaction.
The entire labor-intensive and error-prone process was replaced by an automated intelligent data recognition workflow, which involved no manual intervention. The result was a fast response time and a much more satisfying customer experience.
In this use case, detect intent in emails, learn how we used machine learning and NLP intent analysis to focus on intent detection for customer emails (e.g. a change in policy request) and automated routing to the appropriate person. The result was a response time three times faster for customers.