This project is read-only.


This solution is an Opportunity scoring template designed to predict the likelihood of a win for each sales Opportunity in Dynamics CRM.

The model was built on historical Opportunity data and account related data available in Dynamics CRM out-of-box entities and fields.

In Dynamics CRM, after installation, you will be able to see not only a opportunity win score, but also an explanation of the key attributes that contributed to the Opportunity score.  This is useful to identify potential steps that could improve the Opportunity win score.

Watch video for "Machine Learning on Sales Pipeline"

 You can download similar solution for Sales Force here

Architecture Components 

  • Machine Learning model - "Opportunity score" built in Azure ML
  • Machine Learning model/techniques to provide insight of feature contribution and recommendation
  • Retraining Pipeline to automatically train model on a recurring basis
  • Scoring Pipeline to score model
  • Azure Web jobs for automated / scheduled run
  • Dynamics CRM forms
  • Installer for setting up opportunity scoring for Dynamics CRM
  • Azure Key Vault for storing all confidential information entered by the user. Only the user has access to the Key Vault (Microsoft does not have access to the Key Vault or data stored within the Key Vault) created by her/him in her/his Subscription.


  1. Download zip file. Extract zip file.
  2. Inside the extracted folder look for "CRMInstaller.exe". Double click on this executable.
  3. Follow the wizard.  
  4.  User (using Installer) would need to enter valid credentials for existing Dynamics CRM instance with administration rights & an Azure subscription.
  5. Select the Subscription and Region where you want to install model and provision Azure resources.
  6. You can (optionally) enter an email address and mail server to receive alerts from the retraining and scoring system.
  7. The installer then creates the following required resources, in the selected subscription, for setting up the Opportunity Scoring model
  8. i. Creates an Azure Storage account to be used by the model to store configuration files, logs for the        model run and the monitoring files.
    ii. Provisions and creates an Azure Machine Learning Work space in the selected region which will contain the copy of the Opportunity scoring models to be used for scoring the opportunities.
    iii. Copies the Opportunity Scoring Model experiments to the created Azure ML work space
    iv. Creates Azure web jobs which facilitate the automated run of the Opportunity model, facilitating the  data ingestion from CRM, scoring the opportunities and writing back the scores to CRM. The web jobs  are scheduled to run on a daily basis.
    v. Provisions Azure Key Vault for safe keeping of sensitive data like CRM password, Azure Storage Key  etc. This approach keeps the passwords protected from use by unauthorized use by any other  application or user.
    vi. Creates and uploads a configuration file to the Azure Storage account. The configuration file contains  data required by the Opportunity Model run, such as CRM account URL, list of entities and attributes of  CRM to be used, mail server information etc.
  9. At the end of the successful installation you will get a path of a Text File for reference (with details of all the resources created and associated with the application for CRM Tenant). You can verify and see details of the all the resources created for the CRM Tenant. Log of Installation process to visualize the activities by stages is also generated and path provided in the Text file.



Step by step installation
 Using Opportunity scoring in Dynamics CRM
Machine Learning models
Installer Features

Benefits of opportunity scoring

Useful links

Dynamics CRM trial          :

Pricing                                 :

CRM guided tour               :

CRM Compliance               :

Cortana Intelligence suite:

Last edited Sep 2, 2016 at 7:20 PM by prashdesh, version 38