BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Salesforce IoT Cloud made its debut just over a year ago, and while the platform has grown up since then, companies interested in using it face a lot of trial and error.
Salesforce initially delivered its Internet of Things (IoT) Cloud, with the Thunder development engine underlying it, to ingest massive amounts of data streaming. It has since combined that with slower-moving contextual data about customers. Taken together, that data helps companies make intelligent decisions about their products, services and customers.
For example, real-time data mixed with contextual data in Salesforce IoT Cloud can be used to alert manufacturers to the fact that not only is there an issue with a vehicle part, but also when the vehicle was last serviced, where the vehicle is, whether the car owner is up for a new lease soon and much more, explained Dylan Steele, director of product marketing for Salesforce IoT Cloud and App Cloud, at this year's Dreamforce conference.
"Our whole goal here is not to be reactive to the customer, but proactive," he said.
It is early days of IoT, though, and customers experimenting with IoT Cloud uses might try 100 different versions of an IoT use case before they determine what's most meaningful for their business, Steele said. Part of that is determining what to measure, what to build around, and how to tweak it, test it and optimize it for customers, he explained.
Companies using IoT Cloud, such as Emerson Climate Technologies, which offers connected thermostats, send Salesforce their thermostat health-check data. They are analyzing the data to understand when and why there is an issue, so they can identify it before a system breaks down.
Meanwhile, Salesforce is looking at ways to apply its new Einstein artificial intelligence (AI) technology to IoT Cloud. PredictionIO, one of the AI companies Salesforce acquired, allows companies to run machine-learning algorithms on large data sets. It will be embedded in IoT Cloud to score data for relevancy and provide recommendations to fix problems, Steele explained.
"We are slowly bringing those two technologies together," he said.