The big announcement at Dreamforce 2016 was undoubtably Einstein. Based on a number of acquisitions and internal development, Einstein is Salesforce's first productisation (yes, that's an awful word) of AI technologies. The technical components are covered in a previous post here.
A few components have been liberally mixed with flour, eggs and butter and are baked into the product offering as Salesforce Einstein. There's a lot of Trailhead on the subject and the Salesforce homepage has some nice videos with music that Shazam didn't recognise. Yes, I Shazamed the music to a Salesforce marketing video. Feel free to stop reading now.
The meat and gravy of Einstein is mainly in Sales Cloud which is covered here.
Einstein Activity Capture
Using the argument that salespeople spend too much time on admin and not enough time on selling, activity capture provides functionality that connects Office 365 and Google mail accounts to Salesforce to help relieve that chore. I have to say that I always find the excuse that salespeople spend too much time on admin a particularly weak one as I'm firmly in the camp that admin means a salesperson is well organised. Activity capture then captures relevant emails, appointments and tasks against Salesforce records so the salesperson doesn't have to. There's control over what's added within either the mail client or Salesforce to ensure sensitive or secure (or irrelevant) content isn't automatically added. Reporting then shows stats on tasks, meetings and contacts to see who's most active against their accounts.
Einstein Automated Contacts
Based on email and activity data, Einstein Automated Contacts will either automatically add contacts into Salesforce or ask the user if suggested contacts should be added. List views will allow you to view which contacts have been added by Einstein.
Einstein Lead Scoring
Many Salesforce users add formulas to their leads based on criteria such as job title, location, industry, etc which produces a score that can be used to rank leads. In many Einstein does exactly this. The obvious question therefore springs to mind - why pay for new functionality that a simple formula can solve? The major difference is Einstein's ability to adjust it's lead scoring based on historical data. We think we know which leads are more likely to convert but how often do we perform any kind of analysis to check we're correct? The answer is not often. Based on analysis of historical trends Einstein will automatically adjust it's lead scoring to match the characteristics of leads you're closing rather then the leads you think you're closing.
Einstein Opportunity Insights
In common with Lead Scoring, Opportunity Insights use historical data to help you take appropriate actions at the correct time. There are 3 types of Opportunity Insights;
- Deal predictions. Based on historical activities deal predictions will tell you if an opportunity may/may not close.
- Follow-up Reminders monitor communications between your sales team and their opportunity contacts to let you know there's been no communication or a contact hasn't responded to specific communication.
- Key Moments monitor events within communications such as a contact using a competitor's name in an email.
Einstein Account Insights
I have to admit that I generally remove the newsfeed from the Lightning Homepage. I, nor my customers, tend to find it specific enough to be of much value. That said, multiple Google alerts on customer accounts are generally ignored as more spam email. Account Insights aims to provide more specific and targeted newsfeed data about your accounts from either the homepage in an easily shareable way via email or Chatter. The AI component of Account Insights is the ability to pick keywords from email and provide timely information based on real events that could materially affect an opportunity such as merger and acquisitions. The example of the customer mentioning a competitor means news about that competitor gets posted to the Insights. This makes to easier to have meaningful engagements with customers as your team is better informed on timely information that could impact an opportunity.
It's all about data
It's got to me said that AI requires lots of data. It's hard to make predictions across limited data sets because there's simply not enough to go on. Humans can't do this effectively so why could code? Average AI algorithms across large datasets will probably do a better job that great AI algorithms across limited datasets so please think about how much data you generate before investing in Sales Cloud Einstein. The functions are an add-on to Sales Cloud functionality with a price premium of $50/user/month (I don't have GBP pricing).
For those who want to wait and see how this evolves, there's also a possibility of a trickle effect of the functionality being added into standard capability over time, especially as competitors start to develop similar capabilities. There's nothing better than a competitive market to deliver value to the customer and the CRM market is pretty hot right now so patience may be rewarded, although not at the expense of competitive advantage when the competitor down the street is killing you due to harnessing the benefits of AI.