Einstein, meet Watson: Salesforce and IBM combine forces on AI


Salesforce.com Inc. is improving its Einstein deep learning platform via a new partnership with IBM Corp. that will see Big Blue’s Watson AI business unit integrated with the service.

The partnership, announced Monday, effectively makes IBM a kind of AI-consulting partner for Salesforce, allowing it to sell new services across both Einstein and its Watson platform, most famous for beating two Jeopardy champions in 2011.

The thrust of the announcement is that IBM will make Watson’s insights available directly within the Salesforce Intelligence Customer Success platform. Users will be able to draw on both Einstein’s customer relationship insights and Watson’s enormous trove of data that spans industries including healthcare, financial services, retail and weather forecasting, the companies said.

So, for example, an insurance company using Salesforce’s platform will be able to tap into Watson’s weather data to target customers ahead of an incoming storm, knowing full well those customers are likely to be interested in protecting their property. The partnership will also enable marketers to better target consumers by pairing their Salesforce data with IBM’s local retail trends.

“Within a few years, every major decision — personal or business — will be made with the help of AI and cognitive technologies,” Ginni Rometty (pictured, right), IBM’s chief executive officer, said in a statement. According to Rometty, IBM’s Watson platform will “touch” more than a billion people in the next few years, making an impact on everything from connected cars to tax preparation to oncology.

“The combination of Einstein and Watson will make businesses smarter and our customers more successful,” Marc Benioff (left), chairman and chief executive of Salesforce, said in a statement. “I’m thrilled to form an alliance with IBM — no company’s core values are as close to Salesforce’s as IBM’s. It’s the best of both worlds.”

Einstein: Room for improvement?

The companies are pitching the deal as a straightforward partnership, but in reality it’s probably fairer to say that Salesforce has signed up as an IBM customer. Despite all the hyperbole around Einstein following its launch last September, Salesforce isn’t even remotely close to where IBM Watson is when it comes to AI, Patrick Moorhead, president and principal analyst at Moor Insights & Strategy, told SiliconANGLE.

“Salesforce needs IBM because they’re still way behind in AI,” he said. “Meanwhile, IBM is a leader in it.”

That idea is given further credence by customers’ initial perception of Salesforce’s AI capabilities following the launch of Einstein at last year’s Dreamforce conference. Customers who met with Deutsche Bank analysts on the eve of that event said in a note to the investment bank’s clients that the consensus was that “the Einstein/AI offerings are intriguing but super early-stage, mostly marketing.”

Moorhead said that Einstein was a “good start” for Salesforce in AI, but that when you compare it to where IBM was when it started a decade ago, it’s still a long way behind Big Blue. “I mean, if Einstein were all that, why would Salesforce need to partner with them?”

Salesforce sought to put some flesh on Einstein’s bones with the announcement today of the company’s spring release of its software. Among the new features are Einstein Vision, a set of application programming interfaces that let developers use image recognition in CRM and build AI-powered apps quickly. Salesforce said companies such as Air France-KLM, U.S. Bank and Seagate Technology are already using Einstein.

Question marks also remain about the two companies’ ability to properly integrate Einstein and Watson, Holger Mueller, principal analyst and vice president of Constellation Research Inc, told SiliconANGLE. He also pointed out that Watson has never been used to scale for a Software-as-a-Service vendor before, which should make things “interesting.”

“Watson has shown it works for customers and individual engagement, but not for thousands of Software-as-a-Service customers,” Mueller said. “Also, Watson will be running on IBM’s systems, so it remains to be seen how the necessary data will be made available to it.”

In any case, the companies should be able to iron out these architectural details and with Watson’s smarts, Einstein might finally become a truly useful product. IBM hopes so too, and to that end its Bluewolf consulting firm has announced the formation of a new practice unit for Salesforce Einstein’s software and services to help customers deploy the new solution. Also as part of the deal, IBM said it will begin using Salesforce’s cloud services to organize its own customer support services.

The newly integrated Einstein/Watson offering will be made available to customers in the second half of this year. Pricing details will be announced at a later date.

Image: IBM