Salesforce X AWS Machine Learning
Download as PPTX, PDF1 like454 views
This document outlines the process of predicting salesforce campaign member responses using AWS machine learning. It involves adding members to a campaign, triggering an AWS ML process to analyze member attributes, and generating a model based on the collected data. The final step includes calling an API to obtain predictions on whether members are likely to respond to the campaign.
Salesforce X AWS Machine Learning
- 1. Salesforce Campaign Members Prediction with AWS Machine Learning
- 2. Overview : How to work with AWS Machine Learning 2 DM Campaign to Top Customers 1 3 2 Trigger AWS ML 1. Add members from Contact / Lead records 2. Once members are added to the campaign, Apex trigger passes each member’s attributes to AWS ML to get predictions, who are likely to respond the campaign and who are not 3. The predictions via AWS ML are shown here
- 3. Architecture Trigger DM Campaign to Top Customers Completed campaign records store each campaign member’s status such as sent or responded. csv file Report Create the ML model based on csv file stored in SFDC. The ML model is a collection of patterns that Amazon ML finds in your data during training Campaign members are added and then the trigger is called to execute callout request to AWS API Gateway to get the predictions Feed those data to Machine Learning for prediction
- 4. Prepare Training Data for AWS Machine Learning 4 Campaign Each campaign members (Contact) has specific attributes for ML such as education, job category, and so on Once the campaign is completed, the trigger converts campaign member’s status (sent or responded) along with all attributes on his contact/lead record into AWS Training Datasources record. This example generates three AWS Training Datasource records because there are three campaign members for this campaign. Same data Target 1 : responded Target 2 : sent Contact AWS Training Datasource
- 5. Upload Training Data source to S3 5 1. In order to feed training datasource to AWS Machine Learning, save all records as csv file via report function csv file 2. Upload csv file to S3 backet
- 6. Set up AWS Machine Learning and API 6 1. Create ML Model from csv file on S3 bucket 2. Review the ML Model’s Predictive Performance 3. Create Lambda function to call the ML model 4. Create Rest API to receive parameters and pass them to Lambda function in order to get prediction based on the ML Model 1 2 3 4
- 7. Call the API from Salesforce 7 Once campaign member is added to a campaign, Apex calls the prediction API with the member’s attributes based on his contact record.
- 8. Result 8 The API returns Yes or No which the member is likely to respond the invitation and purchase products.