Salesforce is the most successful SaaS company in the world. AWS is the biggest infrastructure player, and the two giants have gotten together over the years whenever it makes sense. As a couple of examples, the two signed an agreement in 2016 for Salesforce to use AWS infrastructure. In 2018, they got together to make it easier to exchange data between the two platforms.
This year, they are extending the partnership yet again with an announcement this week at Dreamforce in San Francisco. This one involves moving Salesforce data to SageMaker where customers can build machine learning models based on that data. When they’re done, they can move it back to Salesforce where Einstein can use that model to drive intelligent tasks.
The whole process is made possible by Genie, the data integration layer the company announced yesterday.
Rahul Auradkar, EVP and GM for unified data services and Einstein at Salesforce, says certain customers have built workflows in SageMaker and they wanted to make it easy for them to work where they are most comfortable, while taking advantage of the data inside Salesforce to build their models.
He says that it starts with moving the data from the Salesforce CDP into SageMaker using Genie as the data conduit. “From there you build the model and then we meet the data scientists where they are, essentially enabling data scientists to use their familiar tools on SageMaker to build their models. Then, they can bring the models into Einstein and run the inferencing there,” he said.
Liz Miller, an analyst at Constellation Research, says the partnership benefits Salesforce customers in a couple of ways. “For those who have data and analytics teams and have been working on AI and models, this makes it easy to bring those models to Genie and open up those models to the mass of data Genie can hold. This has been a request from customers, especially large enterprise customers, from what Salesforce execs have told me,” she said.
She adds, “For many customers, and we hear this often, models and especially AI models and projects can stall because they can only be let loose on limited data sets or not enough customer data to reach a satisfying level of decision velocity. So this partnership connects the models to the last mile of learning.”
It could also simply involve comfort with SageMaker as a customer’s model building tool of choice, or it could possibly be because the model uses both customer data and other external data. As an example, a healthcare company may use customer data in conjunction with clinical trial data stored in an external repository outside of Salesforce and they want to build the model using both types of data inside SageMaker.
“The SageMaker integration enables these organizations to use custom AI models that leverage real-time customer data from Salesforce and clinical trial research together,” a company spokesperson explained.
While Einstein comes with many intelligent processes such as finding the most likely customer to buy, or conversely, the most likely to churn, there are often going to be customized scenarios that won’t be available out of the box, and being able to move various types of data into a model in SageMaker and then back into Salesforce could prove useful for a number of customers.
The Genie announcement this week will likely lead to similar partnerships over time beyond this one, as other companies look to take advantage of Salesforce data in their workflows.