What features can users expect from Salesforce Einstein AI?

Cloud-based artificial intelligence is still new enough that no one is quite sure what it can really do, or how to integrate it into the enterprise. But Salesforce has had the benefit of learning from IBM's mistakes with Watson, and seems better focused as it moves forward with Salesforce Einstein, its own AI service.

That forward motion involves leveraging Salesforce Einstein AI into an area where the vendor shines -- insights and predictive analytics about customers. Einstein extends Salesforce's existing CRM functionality, essentially serving as a machine learning environment for behavioral data on customers. That, in itself, is formidable, but also somewhat narrow: Salesforce is masterful at descriptive data, but has no real foundation in AI-based problem solving.

Its new partnership with IBM, which basically bundles Salesforce Einstein AI capabilities with the more generalized platform Watson -- of Jeopardy fame -- addresses these shortcomings. It capitalizes on Einstein's strengths: Watson's problem-solving capacity and ability to absorb unstructured big data, combined with Einstein's strong customer behavior metrics result in a powerful engine for insight into customer issues and sentiment.

The market potential of this sort of information isn't yet fully clear, but the possibilities of Salesforce Einstein AI should definitely be watched closely.