<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=1178143905890834&amp;ev=PageView&amp;noscript=1">

QSRs Must Listen Beyond The Orders: An NLP Case Study

Have you ever waited for a cup of coffee longer than you should at a coffee shop? And furiously wrote a one-star review on Yelp.com? How do you know that the coffee shop heard you? Well, there might be a Skellam AI solution that pays attention to every word you have to say about your favorite coffee brand. 

Customers like you and me typically express our sentiments through various channels, for example, direct channels like customer support centre conversations, response to surveys, and product reviews. And on some indirect channels like social media conversations, such as rants, posts, tags and comments. 

Here’s a case-study of how Skellam AI’s solution, an AI enabled text analytics product, helped a Fortune 500 enterprise get insights into customer and employee sentiments.

This solution is a natural language processing (NLP) engine, which is a customised version of our proprietary product, LexCore. It understands unstructured and structured data from multiple sources to deliver insights on the feedback received by a brand.

In case you are wondering how all of this is possible, the NLP engine extracts themes based on conversational tones. Some of the themes could be customer experience, employee behavior, product quality and loyalty management programs. Based on whichever particular segment the business is in, these themes could vary and our solution comes up with them automatically, just like that. 

Our solution detects a theme and then delineates the sentiments for these themes and helps the brand respond or engage accordingly. Using Lexcore, a brand can then isolate negative themes for a particular geography and pay attention to that to improve CX and sales.

The Fortune 500 company was initially using an industry leading social listening tool. But Skellam AI built the application in a few months and surpassed the quality of the existing social listening tool along all KPIs (Key Performance Indicators). 

Presently, Lexcore processes a wider array of inputs including the customer survey responses, customer support calls data, product reviews, social media mentions, earned media, handwritten feedback forms from physical stores (digitized), app based interactions, etc. During the Covid Pandemic lock down, our solution was responsible for gathering a lot of insights into customer behavior and employee morale which enabled our client to make quick operational decisions. Mind you, this was when everyone else had no clue on how to approach the new normal as it were.

Moreover, Lexcore has the ability to detect themes by itself, apart from the predetermined themes set by the analyst. It is also capable of detecting anomalies, where new conversational themes, especially negative ones, spike. This helped our client pay close attention to select new conversations and make corrections where necessary. 

In order to deliver such insights, our solution typically processes 30,000 to 40,000 records from multiple sources in a day. This multiplies into a whopping ballpark figure of 14 million (jaw drop) records in a year or about 1.2 million records per month. 

We are in the process of extending the capabilities to natural language understanding, where a user can directly ask questions and have a response from the product. For example, if an area manager keys in a query such as “Tell me a store with the most number of wait time complaints ,” our solution is designed to give a direct answer to the question based on its AI capability. 

To learn more about the business impact of our solution you can download the case study here.

Don’t hesitate to reach out to us at beawesome@skellam.ai if you’d like to replicate similar magic in your customer engagement and sentiment analysis using NLP.

Download Case Study

About The Author

Lakshmi Narayan V

Content Marketer. Student of Life. Someone with a passion for learning.

Leave a Reply