How Salesforce Is Preparing For The ‘Artificial Intelligence Revolution’
The CRM giant is rolling out its 'Einstein' project, which combines machine learning, predictive analytics, natural language processing, and smart data discovery. Here's what it means for online-to-offline marketing.
Salesforce is unveiling a new research project — ambitiously dubbed Einstein — that seeks to bring the wide-ranging tools associated with artificial intelligence to businesses at all levels.
The project is meant to build on Salesforce’s existing path towards offering predictive analytics, as companies like Google, Apple, and major app developers generally seek to give consumers what they appear to want before they ask for it.
“There’s a revolution going on in artificial intelligence,” said John Ball, SVP and GM of Salesforce Einstein, during press and analyst conference last week. “This revolution is really being driven by the major trends we’re seeing within sort of the last ten to 15 years, starting with web, and the cloud, social, mobile, IOT.
Understanding Consumer Preference In Advance
“Today everyone and everything is connected and all of those connections are generating orders of magnitude of more data than ever before,” Ball continued. “Thanks to the cloud, you basically have unlimited computer storage capacity. It’s really that combination that is driving that revolution in AI.”
Ball cited “various forms of AI already power some of the world’s most popular consumer experiences,” such as Apple’s Siri, which leverages natural language processing to recognize voice commands, on to Facebook’s deep learning facial recognition algorithm that can instantly identify a person with nearly 98 percent accuracy. He also pointed to the use of machine learning by Amazon, Netflix, and Spotify to understand “how each item in their massive catalogs relates to the other and each customer’s preferences.”
With Salesforce Einstein, AI capabilities will be embedded with “every Salesforce Cloud. It leverages all data within Salesforce—customer data; activity data from social media chatter, email, calendar entries, and e-commerce; social data streams such as Tweets and images; and even IoT signals—to train machine learning models.
Considering the scale Salesforce has at its disposal, with millions of users sharing their purchase pattern information daily, the company believes it already has significant wherewithal to automate accurate predictions for its clients and their consumers.
Location At The Center Of IoT And AI
While Salesforce executives didn’t specifically cite geo-data as a factor in Einstein’s predictive insight machine, the fact is clear: location signals underlie all the data that goes into social, mobile, and IoT.
For example, Einstein relies on “image classifiers” that are imbued by place-based attributes to deliver a fuller picture of what a consumers’ needs tend to be at a given moment in time, Richard Socher chief scientist, noted during the demo.
To illustrate his point, he suggested looking at a consumers’ images from a beach.
“Let’s say you wanted to understand where people are using your product, such as a beach towel — you input that to your ‘beach classifier,'” Socher said. “Then, you just click on create and change to add more labels to, which offers a whole host of different kinds of focus to what a consumer might be thinking of. That is just one of the first deep learning phase developer tools that we’ll make accessible [and that will allow clients to] create complex business processes and improve them, with smart, automated decision making.”
The company plans to show off Einstein’s capabilities in full at its next Dreamforce event on October 6.