Here are a list of Winter ’18 Salesforce Einstein features below.

There are additional Winter ’18 features available to Lightning Experience only, both Lightning Experience and Salesforce Classic, Salesforce Classic only or Mobile. Visit the Winter ’18 Release Highlights post.

1. Sales Cloud Einstein enhancements: 

• [Classic] Add the Lead Insights component to your Salesforce Classic lead page layouts so that sales reps see their leads’ scores and the insights that support them.
• [Classic and Lightning Experience] Specify the pattern for Einstein Lead Scoring to use by specifying your lead conversion milestone (lead converts to contact and account or lead converts to opportunity, contact and account). Previously, the default would only analyze historical leads only if sales reps had created opportunities during conversion.  
• [Lighting Experience] (Beta) Allow Einstein to automatically add contacts and contact roles to Salesforce
• [Lightning Experience] (Beta) Salesforce provides a dashboard that shows Einstein Automated Contacts.
• [Lightning Experience] When Einstein Activity Capture is enabled, users can view their Microsoft Office 365 in a “Microsoft Calendar” in Salesforce, automatically added for them.
• [Lightning Experience] You specify how long Einstein Activity Capture data should be retained, up to 5 years worth by contacting Salesforce support. Currently, it is set for 24 months by default.

2. Einstein Discovery identifies ways to improve your story by detecting extreme values, identical values, variables with little variation, and highly correlated variables, to build a model based on best practices.

3. Einstein Discovery allows you to perform comparisons of a time period with the equivalent period in previous years to learn which factors drive changes over time in the Story Timeline.

4. In Partner Central communities, show members which files are popular with the Recommendations component.

5. Einstein Engagement Scoring puts a score on how likely each of your subscribers will open an email, click a link, and remain subscribed.

6. Einstein Vision Version 2.0 has been released.

• When you access the Einstein Platform Services APIs, use the new https://api.einstein.ai endpoint.
• Using feedback API calls, you can correct image misclassification by add that image to the correct label in the dataset, then train the dataset and update the model.
• Use two prebuilt image models (Multi-Label Image Model and Scene Image Model) for predictions right away.

7. (Beta) Einstein Language: Build natural language processing into your apps by analyzing text.

• The Einstein Intent API categorizes unstructured text into user-defined labels for the purposes of determining what users are trying to accomplish.
• The Einstein Sentiment API categorize text into positive, negative, and neutral classes to understand the feeling behind text.