August 08, 2017
Salesforce today launched Einstein Vision for Social Studio, fulfilling a commitment made almost a year ago under the umbrella heading “AI for Everyone.” Einstein Vision uses deep learning to “read” image-based content in social media, meeting an increasingly urgent need as social media becomes an overwhelmingly visual medium, with trillions of photos estimated to be stored or shared online each year.
I spoke with Robert Begg, VP Product Marketing, Social and Advertising Products at Salesforce about the significance of the launch. It shows the leaps that AI has made, he said. “The first thing was sentiment analysis, then automation, then delivering the right content at the right time. This is the next step.”
Efficiency in image recognition does reflect the strides made by deep learning over the last few years. Grasping visual patterns was, until quite recently, something humans did much better than computers. That has been changing, along with the AI-driven capacity to understand and respond to natural language. It’s vital that brands can discover and interpret visual content on social media, Begg said, because users frequently express positive or negative attitudes to brands by posting images, without necessarily referring to the brands by name.
The main use cases for Einstein Vision within Social Studio are:
- Consumer Insights: Learning how people (including) are using products, including the contexts in which the products are shown.
- Brand Tracking: Understanding when people are interacting with brands — for example at sponsored events — based on shared images. (“There’s a ton of money spent trying to activate (sponsored) events,” said Begg. “People might not mention a brand, but will have interacted with it.”)
- Social customer service: Consumers will post visual content on social media reflecting problems with a product, without talking about it.
Brands could — in theory — get to all this content manually: “We often talk about the gee-whiz side of AI,” Begg said. “But it’s also about saving time.”
I asked whether Einstein Vision could be used in Social Studio as a discovery tool, surfacing user-generated content of interest to brands. “That’s not a core use case like some of the others,” Begg responded, “But sure. Listening to influencers to source content could get you to something you might otherwise miss altogether.”
Begg also emphasized the importance for users of having Einstein Vision embedded in the app so that the insights which surface can be immediately activated. Without leaving Social Studio, users can share insights with team members, create cases in service cloud, or enter leads into Sales Cloud. “That’s where you see the real ROI,” said Begg.