Machine Learning & AI

Modern Data Platform

Increasing Customer Engagement through Call Center Sentiment Analysis

A global building materials manufacturer has been at the forefront of its business since the early 1900s. This major developer and producer of construction components employs nearly 20,000 people across the world. Maintaining its foothold as an industry leader is critical. To do so, the company needed fast answers to vital questions:

  • Is there a glut of product quality issues in a certain area?
  • How can we increase responsiveness and engagement with customer needs?
  • Are our rebate programs working?

Azure Speeds Manufacturer Toward Coveted 360-degree Customer ViewAnswering these questions, and many more, could help the company better engage customers as part of its overall digital transformation. By finding a way to monitor the ongoing attitudes of its customers the manufacturer hoped to increase customer loyalty and, ultimately, grow market share and profit. Having an efficient way to gauge product quality, customer satisfaction and watch for potential fraud could also improve call center efficiency.

With between 200 to 500 calls and emails to its customer service departments daily, the company had a wealth of information. However, short of manually listening to and transcribing each call, or reading through every email, there was no way to categorize and analyze the vast quantities of data.

Delving into Customer Sentiment

BlueGranite focuses on creating a great experience for our clients and delivering value at a rapid pace. We teamed up with the company to design a customer sentiment analysis solution. Our goal? To lay the framework for that elusive 360-degree customer view, to foster better customer engagement and understanding. We set out to create an integrated customer view across three different channels – voice, email, and chat. To have this all information in a single repository gives organizations a powerful engagement tool.

Microsoft Azure Makes It Simple

Microsoft Azure makes it easy to piece together different components to create custom solutions. We started with a bucket of information from their email and phone call records, and used a variety of technologies to consolidate the data, transcribe audio, and extract the overall tone of the messages. We employed a variety of Azure technologies in addition to on-premises resources to execute the manufacturer’s vision.

Pulling It All Together

Thanks to multiple Microsoft technologies, the manufacturer has almost instant insight into the challenges and successes of customers across the globe. The client can now easily gauge if they have a happy customer, and, if not, take steps to make their experience better, all thanks to:

  • An automated way to transform audio to text
  • A centralized location to store those conversions, along with email and chat text
  • A system to piece together all of the data
  • An analytics system to monitor sentiment
  • Power BI’s rich visuals to help stakeholders see the big picture

The manufacturer now has almost instant access to between 200 and 500 daily communications and the ability to easily view, report on and quickly respond to customer sentiment. This powerful solution will help the company not just to hold on to the clients it’s got, but also bring new ones into the fold.

Transform Your Enterprise

Ready to make the most of your data? Let BlueGranite help. Contact us today.


  • We used Azure Functions to monitor Azure Blob Storage for new audio files. A new file triggered a speech-to-text transcription from Azure Media Services. The Azure Function also converted the raw TTML output from Media Services into a tab-delimited text file.

  • We employed Azure Data Factory V2 to orchestrate the movement of raw call and email text data through Azure Data Lake Store, Azure Data Lake Analytics, Azure Machine Learning, and Azure SQL Database.

  • We used Azure Data Lake Analytics to obtain sentiment and key phrases from the text, transform the data, and saved the final form to Data Lake Store. We built  custom topic and clustering models in Azure Machine Learning to further enrich and categorize the data.

  • We used Azure SQL Database as a final data mart and Power BI to visualize the data.



Powe[R] BI: Enhance Your
Microsoft Power BI Experience
with R

As a specialized, open source statistical environment, R
represents the primary analysis language for many data
scientists and statisticians. By extending Power BI with R,
Microsoft has created several opportunities to enhance your
Business Intelligence solutions and visualizations.