Technology is evolving every day and people are wondering about the next “big thing”. Business Intelligence is one of those areas which has been under constant evolution and machine learning algorithms are being closely worked upon in the BI segment to move from analytical applications to intelligent solutions.
A recent preview release of Microsoft Cognitive Services and the existing Azure Machine Learning platform has improved the accessibility of building such capabilities into the application.
Microsoft Dynamics CRM is an especially good candidate for using such services so as to improve the way customers run their businesses.
Read this great blog about Microsoft Cognitive Services. It basically provides a collection of intelligent APIs that allow solutions to see, hear, speak, understand, and interpret our needs using natural methods of communication. One such API is Text Analytics.
We decided to test and put this into a real-world use case for one of our customers in the education space.
Our customer is an education institute with thousands of students and uses SolzRBM a Dynamics CRM based solution to manage their core business and students.
One of the most common ways of communication between people is emails and day-to-day email can best reflect the sentiment of the person.
Hence we selected email as our main source of data for this analysis. However, the solution is generic and can be extended to include other communications such as text messages, tweets, etc.
- So, whenever a student sends an email that email flows in Dynamics CRM as an email activity.
- As soon as an email arrives in CRM, the text of the email is read and sent to the text analytics sentiment API in Azure.
- This API then returns a sentiment score based on the email content analysis. Scores are between 0 and 1 (0 represents a bad score and 1 is a happy score).
An average sentiment index is generated for the student (contact) based on individual email scores. If the overall average is
- less than 0.3 then the contact is unhappy
- between 0.3 to 0.75 then average
- above 0.75 then happy
See the gauge dial chart at the student level indicating sentiment index.
This can help and give indications of which students are unhappy and the institute can intervene to address the issues. This may not be 100 % accurate but it can give a great indication. You can even find the overall organization-level sentiment index with this.
Will keep posted on how the results are and how is the organization getting impacted by this.
Till then happy CRMing…
