Artificial intelligence was front and centre all week at Dreamforce, Salesforce’s annual marketing and sales conference in San Francisco.
In the weeks leading up to four-day event, which ended this past Friday, the tech giant laid the groundwork to make a big splash with its foray into AI by announcing its new AI engine, Einstein. And during the event, Salesforce pulled out all the stops to show its belief that AI will shape the future of marketing.
The company splashed branded illustrations of Albert Einstein all over downtown San Francisco, sent an Einstein mascot barreling through presentations, sports game-style, and even had a digitally-rendered animation of the famous scientist “join” CEO Marc Benioff on stage via digital screens during his keynote.
The theatrics and styling may have been over the top, but they signaled just how serious Salesforce is about the power of artificial intelligence. On Thursday, the director of product marketing for Einstein, John Ball, went as far as saying that the industry is “at the beginning of an artificial intelligence revolution.”
To date, Salesforce’s main two AI-powered features for marketers are predictive email and visual listening, but the size of the investment the company has made in AI suggest more is on the way. Salesforce has spent a reported $650 million on purchasing leading AI companies, including BeyondCore, Tempo AI and Implisit Insights. Those acquisitions helped the company build out the team of 175 data scientists that built Einstein.
The company’s pitch for predictive email is that it will make email marketing smarter and, with more automation, simpler for marketers. The tool predicts things like which customers will open emails, which will unsubscribe and which have a high lifetime value. As each interaction with a customer pumps more data into the system, it gets to know the customer better and finesses its approach. The biggest selling point may be that the system is automated, and the company actually claims it works better with fewer rules put in by the marketer.
The predictive email tool was beta tested over the summer by a group of 15 brands, including the online retailer ShopAtHome, which saw a 30% lift in subscribers opening. Another brand in the test group, Aldo, used Salesforce tools to reduced the number of emails it sends by 40% while increasing email revenue by 70% by sending more targeted, relevant messages.
As for the visual listening tool, that feature uses machine learning to identify particular products and brand symbols, then provides reports to the marketer with recommended actions. For example, the tool could spot a logo on a coffee cup posted to Instagram, then suggest the brand’s social media manager reach out to the consumer who posted the photo with an offer or information.
While these types of tools are new to many marketers, AI is already powering many products in the consumer world. AI and AI-like tools run the back end of everything from Apple’s Siri voice recognition app to auto-tagging on Facebook and recommended products on Amazon; all three of which were inspirations for Salesforce as it built Einstein.
Jim Sinai, vice-president of marketing, Einstein, said the goal for Einstein is for consumers not to know it was an AI tool that delivered their email, or, in the future, other types of marketing messages.
“Good AI should function like magic,” Sinai said. “You shouldn’t feel like you’re using AI.”
While Salesforce has placed a big bet on AI, Mark Lush, principal at Deloitte Digital, said most companies are still at the very early stages of understanding AI and how it can help them run their business. In fact, Lush said he sees many of the recommendation features being branded as AI as “predictive analytics on steroids” rather than true artificial intelligence.
“We’re not seeing AI being used in its purest form,” Lush said. “AI is a term that’s not well understood yet. Some people use AI, but what they’re really referring to is predictive analytics. A lot of organisations are using predictive analytics.”
In the coming months, Lush said he expects companies to move from predictive analytics to more automated systems of selling and marketing using AI-engines.
“The technology is maturing rapidly month by month and year by year,” he said. “The adoption and embracing of the technology into the enterprise is really only just about to begin to happen.”
“The possibilities and use cases for AI, most companies are beginning to discover them,” Lush said. “I don’t see a lot of companies using AI deep within the business yet, I think the next two to three years will see a significant growth in that area.”
