If Salesforce is jumping on the AI bandwagon, IBM's Watson is giving it the final shove onboard

Cloudy giant's Einstein sucks on tech behemoth's data hose

Robot talks unto robot photo via Shutterstock

A partnership between IBM and Salesforce involving the namesakes of business and science heroes: Watson and Einstein. It must have sounded like a genius move in the PR brainstorming session.

IBM’s data-processing system Watson is named after Thomas J Watson, the chairman and chief executive of Big Blue back in the day. Salesforce has appropriated Albert Einstein for its fledgling analytics service, announced four months ago. The news of a collaboration between Watson and Einstein was uncorked as Salesforce announced the Spring 2017 release of its CRM, which will, of course, include Einstein.

As is customary in such big announcements, the corporate brass hailed the news in breathless and all-encompassing tones to the highest heavens. Watson will provide an “unprecedented understanding of the customer,” claimed IBM chairman, president and chief executive Ginni Rometty. As IBM eviscerates staff numbers in the name of reinvention, Watson is in Big Blue's basket of initiatives deemed “strategic.” IBM claims Watson makes up 40 per cent of its sales figures, but refuses to actually break out Watson’s revenue numbers.

“Einstein and Watson will make businesses smarter,” said Salesforce’s walking PR machine Marc Benioff. The Salesforce chief was “thrilled” to be entering an alliance with IBM. And therein lay the rub: Benioff should be. The partnership will, basically, pump structured and unstructured data about "weather, healthcare, financial services and retail" from Watson into Einstein in the second half of 2017.

The idea of Einstein is to help Salesforce users make sales and marketing decisions from real-world data provided by Watson. Einstein is, predictably, billed as an artificially intelligent system, sorry, service, no wait, platform. Meanwhile, IBM will deliver an Application Integration Suite for Salesforce, to combine on-premises data with information from its cloud store, and surface that information in Salesforce by the end of March.

Salesforce consulting agency Bluewolf – which IBM bought in May 2016 – will help customers deploy the combined Watson and Einstein capabilities through a new unit. Bluewolf will develop industry-specific packages to help uptake, also in the second half of the year.

Unveiled in September 2016, Einstein provides machine learning, predictive analytics, and data discovery for the Salesforce CRM. At the heart of the software is PredictionIO, an upstart Salesforce bought in early 2016 and whose open-source framework was handed over to the Apache Software Foundation (ASF) and turned into an incubator project.

That project has been chugging along for a bit, with more than 4,000 commits and 113 contributors, although development has slowed down – presumably while Salesforce concentrates on crafting the code for its commercial Einstein service.

Benioff reckons Salesforce is on track to become the fastest software company to hit $10bn in revenues. One way to do that is to jump on the machine-learning bandwagon, with IBM giving Einstein a final shove onto the cart. If open-sourcing PredictionIO

Incidentally IBM will deploy Salesforce’s Service Cloud. ®


Biting the hand that feeds IT © 1998–2017