Salesforce is celebrating the one year anniversary of unveiling the Einstein artificial intelligence capabilities for its products with a new set of three features for users of its Sales Cloud services.

A new Einstein Forecasting feature will use machine learning to help sales managers provide more accurate predictions about how well their team will perform over the course of a given period, based on all the data stored in Salesforce.

It’s designed to help managers do away with guesswork when it comes to projecting sales revenue and instead use projections based on data science that takes a company’s entire history into account.

Sales reps also get a couple new features with this release. Einstein Opportunity Scoring will provide them with information about which in-progress sales deals are most likely to succeed, along with those that are at risk of falling through. That information should help reps figure out where they should be spending time in order to have the most impact.

Einstein Email Insights uses natural language processing to parse the contents of sales reps’ emails and tell them what they need to respond to, plus why that particular message is important. At least in theory, that should help people stay on top of key messages, especially after a day full of customer meetings that could lead to an overflowing inbox.

All three of those features are currently in pilot, and Salesforce expects to roll them out early next year. The company hasn’t disclosed pricing yet, saying that prices will be made available when these features launch.

In addition to all the product news, Salesforce revealed a dedicated $50 million fund for AI companies. Highspot, Squirro, and TalkIQ are the first firms to work with the new fund. On top of that, Salesforce Ventures also invested in All Turtles, the startup studio focused on creating AI companies that was founded by former Evernote CEO Phil Libin, along with Jessica Collier and Jon Cifuentes.