5 Cool Einstein Features to look for in Salesforce Winter ’20 Release

Salesforce Winter ’20 release is just around the corner and I’m pretty sure that like me many of you must be waiting to try out the innovative features that Salesforce will be making available to its customers.

While there is a long list of exciting new features and enhancements in the release notes, the one I’m most excited about are the upcoming Einstein features and enhancements.

Salesforce Einstein, which brings the power of AI to Salesforce, has been making CRM processes smarter for the last 3 years and this year as well it promises to raise the bar with a number of new built-in intelligence and platform features.

Here is a list of top 5 Einstein features in the upcoming release that will make our CRM systems smarter and will provide capabilities to the Einstein platform developers for building intelligent solutions:

Let us look at them one by one:

1. Bring your own NLP into Einstein Bots

The feature will expand the horizon of Einstein Bots by allowing customers to connect the Einstein Bots to their own natural language processors. This will enable Einstein Bots to handle complex NLP scenarios which cannot be handled by the Salesforce’s out-of-the-box intent sets. The capability will also enable customers to process Bot utterances in multiple languages by plugging in the custom NLP, whether you are using Microsoft Azure, IBM Watson or AWS NLP. While the out of the box Einstein Language models can deliver exceptional customer service, the addition of this feature will help customers to deliver bespoke Bot experiences through Einstein Bots.

2. Einstein Article Recommendations

Knowledgebase articles have been the best buddies for customer service agents for ages and have been helping agents resolve customer cases through documented procedures and solutions. With Winter ’20 release Salesforce has made the knowledgebase articles even better by enabling intelligent article recommendations using Einstein AI. The feature provides the ability to train the recommendation model based on the fields that are most important for the case. Once trained the Einstein Knowledge Article Recommendation lightning

3. Automated Story Set Up and Descriptive-Only Insights in Einstein Discovery

Einstein Discovery has become smarter in the new release it can now choose columns in the dataset that have the most impact on the outcome variable. Einstein Discovery not only increases the speed of the story creation but also ensures that you don’t miss out on the important columns that can help achieve the goal of the story. The Descriptive-Only Insights in Einstein Discovery further speeds up the story creation if the desired output is to just find out what happened without analyzing the dataset for predictions and improvements. The time to analyze the dataset will be reduced with this feature and it will help in faster delivery of the “What Happened” analysis.

4. Conversational Search

Conversational Search feature in this release is a step towards querying of Salesforce object data using natural language statements. Users can get to a filtered list of records by using natural language statements related to the object and the required filters in a descriptive way as compared to creation of list views or reports which take time and effort. The dynamic nature Conversational Search feature doesn’t required the user to navigate to different places in the application to get to the records they are looking for, instead they can simply type the query in the global search box and Einstein takes care of the rest.

5. Identify Biases in Einstein Discovery

Identifying biased variables and removing them from a model is a challenge that needs addressing at every stage of the model’s life cycles. Biases can make their way into a model through different known and unknown factors and can result in a model that gives results that are biased towards a particular gender, age group or ethnicity. The new Einstein Discovery feature is a great step towards identification of such biases in the model. It helps in identification of variables that may be leading to a biased prediction by providing warnings about the biases and what their impacts are along with an option to remove the bias.

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Show Buttons
Hide Buttons