Contextual Language Analytics to improve Customer Service for a Delivery Platform

The demand for a better customer experience is ever increasing. Be it retail or e-commerce, companies today are looking to make every stage of the buyer’s journey the best it can be. Without a shadow of a doubt voice of the customer (VOC) data has become the single source of truth, if you may.

VOC offers valuable and deep insights into what the customers are actually looking for, they preferences in terms of likes and dislikes with respect to your current offerings.

The case is the same even when you take a food delivery platform. They get millions of reviews regarding restaurants on their platform, the service from their platform, delays, bad experiences, good experiences, so on and so forth.

So, does this open up pandora’s box? Nope! But it certainly does open up a gold mine of information that can be used to make the platform and service the best it can be. What better way is there to make your platform the best, than listening to what your users have to say about it!

Enter Text Analytics

Today, customers change their mind about a service, online platform or product faster than you can blink. That being said, VOC is the most valuable tool in your arsenal. We now get into how VOC can be done; the answer is simple – Text Analytics! With a product such as teX.ai, VOC analysis has proven to be quite easy. The few benefits that teX.ai can offer through VOC analysis are:

  • Get into the mind of your customer
  • Discover new trends
  • Stay ahead of the competition
  • Win back dissatisfied customers
  • Know where you can improve your business

The Use Case

Recently, one of India’s largest online food ordering and delivery platforms commissioned teX.ai to perform contextual language analytics to improve customer service. teX.ai had to analyse the problems mentioned by the customers in the chat and derive insights from it, in a timely manner.

Given the large volume of traffic and numerous customers using the platform, it is critical to ensure proper and effective customer outreach and customer service. One of the major mediums used by the client to effectively communicate to their customer is using the in-app chat functionality.

The Requirement

The client had access to all the chat messages from their customers describing their issues while using the app or the quality of the food. The client’s interest was to analyse the problems mentioned by the customers in the chat and derive insights from it, in a timely manner.

The Objective and the Challenges

The objective was to improve customer service by effectively deriving insights from chat messages to appropriately address customer issues, leveraging text analytics.

There were a few challenges that had to be overcome in order to meet this objective:

  • Processing large volumes of unstructured data, in the form of chat messages.
  • Understanding the prominent keywords and correlating with the various categories and sub-subcategories.
  • Timeliness to complete the entire process – from receiving the chat message to acquiring the insights to addressing the issue.

Solution Time

Technically Speaking

Text Analytics Model

Text analytics models like ELMo were used to start processing the messages to find various patterns emerging from the chats.

Semantic Search

The client provided the chat messages grouped as category and sub-categories. Semantic search was deployed to allow for a streamlined search of queries and specific results for further analysis.

Analytic Snapshot of Messages

Semantic Sentence Similarities using ELMO method was deployed to cluster the similar topics in each of the 9 categories. This presented the output by clustering similar keywords as colour coded data points. The client was able to quickly view the analytic snapshot of their chat messages in a scatter plot, allowing for rapid analysis.

In a nutshell

  • ai was leveraged to analyse thousands of chat messages contextually and provided focused insights to target the problem and start fixing the customer service-oriented problems.
  • The insights were then used for further root cause analysis to avoid its occurrence in the future.
  • We then created a pipeline to process future chat messages and get the required insights in an automated flow.
demo-cta-blog

The teX.ai impact and a happy client!

  • A streamlined and automated process was in place to capture important keywords from incoming chat messages and proper classification for issue redressal.
  • The chat representative understood the chat sentiment with real-time chat insights from incoming messages.
  • Achieved an accuracy level of 80% in identifying customer’s problem area and accurately classifying them.
  • 20% reduction in same type complaints from customers within 3 months, improving customer satisfaction.

To Summarize

This is one particular use case which I have used to elaborate on how text analytics is crucial to analyse and listen to the Voice of Your Customers! With a product such as teX.ai, the use cases are limitless and can cut across any and all domains which operate in the B2C space. It is imperative that text analytics solutions are no more a luxury, but a necessity to stay ahead of the competition.

Author: Abhimanyu Sundar
Abhimanyu is a sportsman, an avid reader with a massive interest in sports. He is passionate about digital marketing and loves discussions about Big Data.

Leave a Reply

Shares