Infosys experts share their views on how digital is significantly impacting enterprises and consumers by redefining experiences, simplifying processes and pushing collaborative innovation to new levels

« Uncover the potential of NLP in SCM & EAM | Main | Empowering financial services with Salesforce Financial service cloud »

Machine Learning Implementation in CRM

Customer Relationship Management (CRM) is the most common platform used in every industry across the world to manage customer relationship and interaction in order to provide best service to their customers and improvise their business. The volume of data streams in CRM applications would be very high as it captures growing data in several stages of the CRM life cycle such as marketing, Sales and Customer service stages.

Most of the CRM packages in the present market provide only to generate reports and BI charts with the available support of their system (out of the box support), these reports and graphical representations can be generated from various tables and columns of CRM database, in addition it also provides graphical representation in the form of pie, bar charts and comes with capability to illustrate various trending information. There is always a need for manual intervention, may be sales expert or external BI tools would need to be employed for forecasting the business and that too, I daresay with only about 50% probability that forecast and actuals would match. Now, it's time to switch to Machine Learning (ML) to improvise your business with rapid and closer predictions to handle large amount of data ranging from Terabytes to Exabytes.

While ML is now proving that it can help predict future prospects based on past data, CRM merely focuses on past and present data and provides insights about customer and sales patterns. However, Machine Learning does continuous learning and provides real-time insights also it provides recommendations over customer, sales and prospects for the best outcomes.

Let's see, how traditional system works and how this can be improvised with ML. For instance, consider the following data from opportunity records,


The above set of data (table-1), probably won't give any insights for the reason behind loss of the business even with existing customer, now look at the below table (table-2). It would give some inference for winning and losing business by just adding more meaningful columns from other tables. This kind of intelligent and real-time data extracts from various tables can be generated and reported in present CRM system with available tools and techniques and this could be extended to some graphical representation. If you are looking even beyond such reports, you need to look no further than ML.


Make a prediction with machine learning algorithm

Now, let's understand about Machine Learning and how this can be used in CRM world. For convenience, lets convert the string into 0s and 1s in table-2 and will apply simple decision tree algorithm to understand machine learning prediction here, then the data representation becomes as given below format (Table-3) and a model (algorithm) to be trained with training data(features) to predict upcoming opportunity record in the CRM system i.e. whether the opportunity has just created in the system is potential or not? (Hot or Cold Deal). We will use Python in this example.


The following Python code snippet illustrates the model and decision tree is a classifier used to predict the input data.



Let's feed the input to the model and see the result, input is just created opportunity record



Now the prediction is completed and we have received a data-driven result. There are even optimization techniques to improve the results to be more accurate. The more we train the model with training data, the more accuracy would be obtained.

As ML always deals with large volume of data, it would definitely help your business with more accurate results when compared to BI tools or any other solutions available. May be in near future, CRM systems will come with out of the box feature in-built Machine learning classifiers for future forecast, in fact all this may be available as a package as part of your Cloud platform.

Apart from insights over customer, sales trend and opportunities and its outcomes, Machine Learning can also be used for automation in customer service, provides suggestion over sales and service, chatbots, feedback etc.

We will discuss more about CRM and Machine Learning in subsequent blogs along with some case study.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.