Sergey Nivens - Fotolia
Though Microsoft has been at it for a while, Salesforce may win the artificial intelligence market war with Einstein, which is well integrated with its other products and easy to use.
As the artificial intelligence market begins to take shape, there are a few major players with different approaches to developing AI for the enterprise.
Their philosophies are markedly different, says Scott Robinson, a BI and SharePoint expert. IBM has long been cultivating Watson, a supercomputer. But IBM's flavor of artificial intelligence remains somewhat academic and removed from everyday human requirements. Microsoft, too, has been developing AI for some time, and its version of artificial intelligence is now infused throughout its applications, services and infrastructure, as it outlined at the recent Microsoft Ignite conference. Salesforce also recently unveiled Einstein, its artificial intelligence technology, which it will bring to all its major clouds, including Sales, Marketing, Service, IoT (internet of things) and, its recent addition, Commerce Cloud. The question has become whether Salesforce Einstein is more enterprise-ready, more development-friendly and more integrated into existing business processes than Microsoft AI.
Microsoft AI technology aims to make workers' tasks more efficient, including organizing email and meetings and finding relevant content. Salesforce, too, is trying to build smart inboxes for sales, help marketers target the right audiences and use AI to make service agents more efficient as they handle a case. But Microsoft, IBM and Salesforce have different approaches. Their architectures are quite different. Robinson said that Salesforce's development platform is more open, integrated and enterprise-ready to make AI easy to use and ubiquitous.
"Salesforce has a very linear technology," Robinson said. "They have been building on the same platform for a while now. The way they do development on this platform is very clean and simple; whereas Microsoft and IBM have built things off the shelf and they have had a hard time having a unified architecture. I think Einstein is going to be a well-integrated, easy-to-use functionality. It will be pervasive. They will be able to attach easily to the stuff they already have deployed."
As a result, Robinson said, Salesforce Einstein could be better integrated into business process than Microsoft AI. While Microsoft has been at work on it for some time, Salesforce Einstein is now integrated in sales, marketing and customer service processes.
"What Salesforce is doing is trying to tear down silos and come up with global data architectures that mesh well with AI," Robinson observed. "Salesforce is able to do this because they grew outward in a controlled and systematic way from their original sales products, and now you can do just about anything on the Salesforce platform that has to do with your business. Microsoft went the long way around. It had this wonderful set of desktop tools that were not specific to any one area of business, and they have tried to do their unification from the bottom up, whereas Salesforce has done it from the inside out."
Microsoft AI pervades its technology stack
Will Microsoft tackle location-based services next?
Salesforce Einstein AI emerges at Dreamforce 2016
Salesforce unveils Einstein Vision for Social Studio