Besides the flying car, few science fiction staples are as long-overdue as Artificial Intelligence. While it’s been a couple years since we saw IBM’s Watson dominating on Jeopardy, AI seems conspicuously absent from our everyday lives, even if there’s a lot going on behind the scenes. With the name alone, Salesforce has set some pretty lofty goals for Einstein, but how capable is it now? Is Einstein’s “Jeopardy” moment imminent, or should we even expect one?
Note: We recently started using Einstein here at Shift (now Silverline)! While this post is a little more future-focused, we’ll be talking about how we’re using it internally very soon.
Watson meets Einstein: What do we know?
We don’t know too much about the recently announced Salesforce-IBM AI alliance, but getting Watson in front of Salesforce’s huge audience, while at the same time giving Einstein an early feature boost until it becomes more mature.
As for features we can expect, natural language recognition is a strong possibility. Even though Watson isn’t actually a huge leap in understanding, relying on fairly standard techniques to analyze questions, the speed at which it can apply these techniques and craft responses is unmatched. It’s likely that we’ll see the most impressive use cases in language-intensive areas like email support before anything else. Being able to intelligently route tickets based on the type of issue, or even prioritize them if the customer seems particularly frustrated, is certainly well-within Watson’s current abilities.
Still, the alliance has some predicting a closer relationship later on. It’s not unusual for IBM to use partnerships to push its technology into new products, even if it plans to release a competing offering, but joining two major AI initiatives this early in development suggests something deeper. IBM acquired a major Salesforce consulting partner last year, and its services division is pushing Salesforce harder than ever in markets like Europe and South America. While we won’t see the fruits of the Watson partnership n actual products for a few years, IBM’s interest in Salesforce clearly extends beyond AI.
Vision: The ‘wow’ feature you won’t see (yet)
Vision is an early indicator of Einstein’s potential to bring ‘magical’ AI to the mainstream. Einstein Vision can be trained to recognize everything from brands to products to specific variations in buildings, but the keyword here is “trained.” Einstein has been pitched as automatic and effortless, so it makes sense that Salesforce isn’t bragging more about a feature that’s limited to developers for the time being.
For now, Vision is quietly being tested in some interesting use cases, such as qualifying leads based on the solar potential of their homes. Using Google Street View and satellite images, it can determine how likely the prospect is to be a candidate for solar based on the angle and area of their roof, saving hours of work and thousands in travel costs. If automatic lead scoring doesn’t impress you, one ambitious developer discovered that it’s quite good at recognizing dog breeds as well.
Is really AI everywhere, or just the new buzzword to beat?
Much like every consumer product needs to be “smart,” every company wants to tell you how they’re using AI. H&R Block seems to think it’s an integral part of doing your taxes (and bought their first Superbowl ad in 8 years to tell us why), but it’s unlikely that most consumers are seeing the impact of AI in their refund checks. IBM itself has been pouring ad dollars into educating consumers on Watson’s usefulness for wine production and airplane maintenance.
Most people do unknowingly see AI on a daily basis, especially if you’re getting directions from your phone or read a news article (that may have been written by a machine). There are already a few handy Einstein-powered features in Salesforce you can try now. Predictive content [link] is particularly exciting, with obvious benefits for sales and support. Intelligently showing products to people based on history is nothing new if you use Amazon, and now you can bring that same functionality to your own online store.
While it can sometimes seem like Einstein and Watson aren’t much more than buzzwords, they’re both trying to solve the same problem. In a world where AI is already being used extensively, how can we put a consumer-friendly face on such an abstract technology?
Einstein won’t be appearing on a game show any time soon, and that’s probably for the best. If you’re familiar with Salesforce, you’ve probably heard of their “clicks, not code” philosophy, and so far that’s been the goal for Einstein; not necessarily to be the most advanced AI around, but to be the easiest to implement for advanced use cases. The real test will be how users react to their new teammate. Einstein-powered features are being deployed rapidly across the platform, meaning that in the coming years, almost everyone who uses Salesforce will have the opportunity to see him in action.
But what will Einstein look like a decade, or even five years from now? Will it be a separate cloud, a licensable technology, or a tightly integrated part of the Salesforce platform? No matter what their immediate goals are, the AI arms race is only just beginning. We may look back on Einstein’s early days as the birth of an essential business technology, and perhaps one that’s every bit as exciting as its sci-fi counterparts.