When it Comes to Einstein, it is All About the Use Case
Day 2 of Dreamforce 2019 saw Marco Casalaina, the VP of Products for Salesforce Einstein, begin today’s AI keynote by making a joke about a prop.
The prop was a small plastic version of the familiar Salesforce Einstein cartoon. A foot and a half tall, colored bright blue, with a shock of messy white hair standing on end, the plastic figure is not like most props. It can speak and even give intelligent answers to complex questions.
Although Marco is a heavy hitter in the World of Salesforce, he was upstaged in his keynote by the small plastic figure. Plastic Einstein has been the star of this year’s Dreamforce because he represents both the promise and the limitations of AI at Salesforce.
Artificial Intelligence is arrayed along a long axis with general AI at one end, and narrow AI at the other. The plastic Einstein is what we would like AI to be – general. General AI promises an oracle like entity that can answer all sorts of questions and solve all kinds of problems using superhuman machine intelligence. That, anyway, is the promise. The reality is that the advances in AI have been in narrow AI – applying mathematically intensive algorithms to tiny problems. This extreme focus is the current limitation of the art.
Dreamforce 2019 – The Promise of Einstein’s Evolution
Last year at Dreamforce, Salesforce introduced Einstein Voice Assistant, allowing sales reps to dictate to Salesforce. With Einstein Voice Assistant, routine CRM transactions such as creating or updating customer records, requesting personalized briefings, or finding dashboards – can now be done through voice recognition.
Now, at Dreamforce 2019, Salesforce promises it will expand the capabilities of Einstein Voice Assistant for any role in any industry. To accomplish this, Salesforce has opened the hood with Einstein Voice Skills to allow developers and power admins to create custom voice apps.
The only way for them to do this successfully will be to define the use cases as narrowly as possible. An AI use case is a specific way an AI algorithm can be used to benefit a business process. Some examples include:
- Predicting the likelihood of a lead converting to a sale,
- Answering routine customer service questions with chatbots,
- Recommending products to customers similar to ones they’ve already purchased or viewed.
Salesforce even offers an Einstein’s Guide to AI Use Cases to give pioneers a step-by-step process for picking a limited use case. The process is heavily framed around a few key questions:
- Which KPIs matter most to your business?
- Which use cases positively affect those KPIs?
- What other companies have been successful with that given use case?
From there, it becomes a question of overcoming three adoption hurdles:
- Identifying a viable use case and having the data to support it,
- Addressing change within employee workflows and embracing the new technology,
- Checking and double-checking the results to see if the algorithm is correct.
It will not be easy, but for ambitious admins and developers, success offers a tremendous career boost. According to Salesforce, 84% of enterprise executives believe AI will help them create a competitive advantage. Still, only 38% of those same companies have an AI strategy today.
Meet us to learn more
Jitterbit customers and prospects will be pleased to know that our company is already well down the path of defining those use cases that matter most to our customers. Please stop by our booth at the #DotOrg lodge at Westin St. Francis hotel or our Integration Concierge booth at the #Trailhead zone to find out more.