In 2016 I was in San Francisco to attend Dreamforce when the company announced the launch of its AI initiative, Einstein. Though Salesforce made a big splash about the announcement, the details were vague and frankly many, myself included, left unimpressed. IBM was making noise about Watson and the Einstein launch came across as trying, somewhat unsuccessfully, to mimic it. However, in the two and a half years since the launch of Einstein, what seemed like AI smoke and mirrors appears to have morphed into something substantial. Recently I was briefed by Salesforce on an upcoming Einstein announcement and had a chance to take a deeper look into where Einstein is today.
The first thing to note is Salesforce is no longer trying to compete with IBM Watson. Rather than leading with AI as a broad solution, Salesforce has baked Einstein into its core offerings very tactically. Instead of touting complexity and power, Salesforce is delivering its AI, with easy to use interfaces and rather than requiring a specialist data scientist to set it up, Einstein can be configured by Salesforce administrators. In the same vein, the outputs of the 35 different AI modules that Einstein runs today are equally simple to digest and leverage. At Deep Analysis we like the approach of small tactical modules baked into service offerings that are easy to use.
It is important though to note that, in reality, there is no such thing as Einstein, because there is no one big AI Salesforce machine. Salesforce has built 35 (with more on the way) individual AI modules and branded them all under the Einstein banner. Some modules leverage Salesforce proprietary AI, others have been built using services from Amazon AWS, Google GCP, and Baidu. Some of the modules are technically simple but use advanced analytics, others are more technically sophisticated and leverage neural networks. SFDC uses an approach of building specific tools to meet specific needs, rather than a blanket, one size fits all approach.
Recently Salesforce announced a slew of new Einstein initiatives, making it more developer friendly and expanding the predictive capabilities with the launch of Einstein Prediction Builder. But what caught our interest was the launch of Einstein OCR (Optical Character Recognition) built on computer vision technology, as it follows Amazon’s release of a similar service called Textract in late 2018.
We have written about this before, but the world of capture is being rediscovered by some vendors and reinvented by others. Mountains of unstructured content, be it in the form of contracts, letters, forms or business cards remain business critical, yet seldom see much love. First generation OCR systems of capture are commonplace but are unloved, underutilized and inefficient. The use of AI here to transform traditional capture is unsexy for sure, it doesn’t have the sizzle of predicting new market trends or segmenting likely buyers from tire kickers, but it is nonetheless an essential element of how many organizations run. It’s too early to tell how well Einstein OCR works or will fare, but it should be noted that Amazon Textract is being heavily tested in the market by ECM vendors and system integrators and is starting to make waves, so the interest in new approaches to OCR is real. Yet managing and leveraging unstructured content has never really been Salesforce forte, hence our interest and it is something that we will watch with interest over the coming years.
All in all, it is fair to say that Salesforce has pragmatically come a long way with Einstein. Bundling the capabilities into the core offering and making them simple to use is smart. How much people actually use them, and how much value these services will truly deliver is more of an open question, as it is for any new technology or service. But to give credit where it is due, Salesforce is making AI accessible to its customers and hiding much if not most of the complexity from them. That’s an approach that IBM Watson may have been better taking, as winning a game of Jeopardy may well be impressive, but it doesn’t tell us that Watson can actually make a difference by converting more leads or reducing operational costs.
Work with us today to ensure you are a disruptor not one of the disrupted!
Get trusted advice and technology insights for your business from the experts at Deep Analysis. [email protected]