≈

AI for Admins: How to Choose the Right Einstein Builder for Your Business Needs

By

Salesforce Einstein was launched in 2016 with a mission to deliver artificial intelligence (AI) into the hands of admins, developers, and business users, helping make Salesforce the world’s smartest CRM. Since then, the Einstein Platform has grown exponentially. Today, it powers more than 116 billion predictions per day for companies of all sizes across the globe!

From some of the largest financial institutions to up-and-coming innovative startups, and from some of the most admired brands in the world to highly impactful non-profit organizations, Salesforce customers everywhere are reaping the benefits of the Einstein Platform. They deliver Einstein’s AI and Machine Learning (ML)-powered predictions and insights to business users and front-line workers in a scalable, secure, and trusted environment.

The Einstein portfolio includes prebuilt or templated solutions designed to work out-of-the-box with Salesforce clouds, such as Einstein for Sales, Einstein for Service, Marketing Cloud Einstein, and Commerce Cloud Einstein. (Learn more about how Einstein boosts each cloud in this Einstein capabilities cheatsheet.)

But what if you want to build your own predictive models, tailored to your data and business needs? One of the greatest advantages of the Einstein Platform is that it enables users to build and automate their own predictions, all with clicks and without the need to write any code. The Einstein Platform offers enterprise-grade capabilities to address a broad range of business use cases. Certain Einstein products are better suited for particular use cases. Which Einstein product can best address your unique solution requirements? Let’s dive deeper into four key elements of the platform so you can choose which one to use and get started with the power of “no-code AI.”

Einstein Prediction Builder

If you’re a Salesforce Admin and you’re looking to turbo-charge your Salesforce Apps with the predictive power of AI, Einstein Prediction Builder is a great place to start. You can use Einstein Prediction Builder to take data from one of your standard or custom Salesforce objects and build a prediction. You also get a Data Checker to determine whether you have enough records to effectively build the predictive model.

Einstein Prediction Builder start screen.

After choosing the object, you can select which records you want Einstein to learn from (example set) and which records Einstein will make predictions for. Simple, right? No mention of behind-the-scenes algorithms here! So, what kind of things can you predict with Einstein Prediction Builder?

Use cases

Einstein Prediction Builder can predict numeric fields, such as a Sales Amount or the days it will take to close a sales Opportunity or a service Case. It can also predict a yes/no outcome, such as the likelihood of a customer buying one of your products.

After building a prediction, you get a scorecard to review the quality of the model so you can confidently deploy to your org. Deploying enables predictions to be surfaced in each record of the object that the predictive model was built on. These predictions can be made available to every CRM user across the org, so everyone gets AI-powered insights, right where they work.

Example Score Card in Einstein Prediction Builder.

Ready to build your first prediction? Check out Einstein Prediction Builder Help or this Step-by-Step Guide. Or, try it in a playground by following along with this Trailhead Quick Start. And the best news? As of the Spring ’20 Release, every Enterprise Edition and Unlimited Edition org includes one free Einstein Prediction Builder prediction, so you can start right away!

Einstein Discovery

We all know that data is what fuels predictive models. What if you want Einstein to learn and build predictions out of data stored in multiple Salesforce objects? Or to integrate data that lives outside of Salesforce altogether? And wouldn’t it be great to get a narrated story that explains the patterns in your data and helps identify the predictive factors that contribute to business outcomes? Enter Einstein Discovery.

A connected data platform

Einstein Discovery is included with Tableau CRM Plus, a comprehensive analytics architecture that enables you to prepare and store billions of records for analysis. Tableau CRM provides a powerful Extract, Load, Transform (ELT) tool called Data Prep Recipes. This tool enables you to build and automate data pipelines that consolidate and prepare data so that it’s optimized for Einstein’s analysis.

Flow chart showing how the data course connects, syncs, and prepares a data set.

Use cases: Similar yet unique

Once you’ve populated your dataset, use Einstein Discovery to create a story (analysis) that focuses on maximizing or minimizing a business outcome, such as a Key Performance Indicator (KPI). Similar to Einstein Prediction Builder, Einstein Discovery creates predictions for numerical or binary business outcomes, such as the likelihood of a customer attriting or buying a product. However, Einstein Discovery also generates descriptive, diagnostic, and comparative insights. Insights include charts and automated, narrated analyses of the data used to train the model. Insights surface and explain significant patterns in your data, enabling you to discern which variables contribute to the KPI you’re analyzing. Although both Einstein Prediction Builder and Einstein Discovery provide numerical and binary classification predictions as well as explanations, Einstein Discovery provides a more thorough, in-depth analysis to help admins understand what’s going on behind the scenes with their data.

Einstein Discovery Story home page.

Analysis created through Einstein Discovery.

Predictive and prescriptive analytics

In addition to predictions, Einstein Discovery models provide prescriptive analysis, called improvements, to help users consider actions they can take to improve the predicted outcome. Users can click the Predictions tab on the story toolbar to explore ‘what-if’ simulations using the model.

Example of Einstein Predictions, Top Improvements, and a Model Overview.

Users select different combinations of features in the interactive panel on the left side of the screen. Based on the selections and the model, Einstein derives predictions and then displays the results on the right side of the screen. To simulate improvements, the user selects an Actionable Variable, which is something they can do to improve the predicted outcome, such as offering a discount coupon or setting up a meeting.

Where you work

The full power of predictive models comes from distributing insights, predictions, and improvements to front-line workers and management alike. Einstein Discovery is great at getting predictions into the hands of users, wherever the users do their work, whenever they need to make decisions. Predictions can be delivered to Salesforce users and Salesforce’s automation tools, but they can also be delivered to other environments, such as Tableau dashboards or even external systems via Einstein Discovery’s API. Check out this blog post for a complete guide to Einstein Discovery’s capabilities.

Example of a set of predictions that can be delivered to Salesforce users and Salesforce's automation tools.

Want to get started? Check out this Einstein Discovery Trail to get your own trial account.

Einstein Next Best Action

So far, I’ve talked about predicting a number, such as the Sales Amount expected for an Opportunity, or the likelihood of an event, such as the Risk of a Customer Attriting. But what if you wanted to recommend an action to decrease the attrition risk based on a range of different possible actions? Or recommend an improvement action based on a business rule (and not only based on past successful actions taken like what Einstein Discovery offers)? It’s time to talk about Einstein Next Best Action (NBA).

A Salesforce org with a tab in the right-hand corner displaying Next Best Action.

Use cases

Imagine you work in a call center where you answer all types of questions and inquiries from customers. Some of them might be looking into buying your newest product, while others might be having a poor experience and asking for help troubleshooting a problem, or calling back to check on the status of an existing claim. You wouldn’t recommend a visit from the technician to the customer looking at buying a new product, just as you wouldn’t offer a new product to a customer who is currently having problems with what they bought from you. Now think about all the possible customer scenarios and potential offers and actions that can be taken to improve a customer’s experience, all depending on what the customer needs. It’s a lot!

Einstein NBA instantly evaluates the context of the situation the customer is experiencing—taking information from the Account or Case object (or any other Salesforce object)—and matches them with appropriate recommendations for what should be offered next. In this way, NBA removes the guesswork that CRM users often face when working directly with customers.

Strategic

Einstein NBA uses decisions and business logic to surface context-specific offers and actions. These sets of potential recommendations and rules are called strategies. Each strategy evaluates all the possible recommendations loaded into it and, after applying the defined business logic and rules, will only surface the recommendations that meet the specified criteria. This saves a lot of time for users, who will now get the right recommendations at the right time instead of manually searching or guessing what the next best action is for a customer.

Once recommendations have been defined, such as an offer on a new product, a discount coupon, or a premium support process, you can filter what will get displayed with Recommendation Logic nodes. For example, new products will only be offered to customers with healthy credit scores, discount coupons will only be recommended for loyal customers with more than three purchases, and premium support will only be offered to customers who pay on time.

Example of a Next Best Action Strategy Builder flow.

Adding predictions to the mix

A key aspect of Einstein NBA is that it can take predictions as inputs for the recommendation logic, such as the ones generated from Einstein Prediction Builder or Einstein Discovery. This way, you can set to recommend your new product only for customers with a high likelihood to buy (as predicted by Einstein) or offer discount coupons to customers with high risk of attrition (as predicted by Einstein Discovery).

Finally, once your recommendation is accepted by the customer, you can launch a flow directly from the NBA component—so if the customer is ready to buy, you can start the onboarding process right there. Accepted and rejected recommendations can be tracked, so you can transform the actual input from customers into improving your NBA strategies and closing the loop on your predictive models.

Einstein Recommendation Builder

Einstein Recommendation Builder uses a comprehensive ML platform and a point-and-click interface to deliver intelligent recommendations for businesses. It combines Einstein’s predictive capabilities and business logic, enabling you to build smarter recommendations for your customers.

Use cases

Einstein Recommendation Builder supports standard and custom Salesforce objects. You can build simple recommendations for everyday scenarios; for example, the “you might also like” suggestions based on a customer’s taste or their past purchase history. You can also build more complex recommendations based on other customers’ consumption; for example, “trending” or “popular” product suggestions that are used on ecommerce stores.

How Einstein Recommendation Builder works

Einstein Recommendation Builder builds recommendations using data from three Salesforce objects:

  • Recommended Items object—what you’d like to recommend
  • Recipients object—who you are recommending the items to
  • Interactions object—past interactions and with data that Einstein learns from

Einstein Recommendation Builder page.

After you select your objects and define your segmentation settings, the Data Checker checks whether each object has enough data to build your recommendation. If your data is in good health, you can build. You’ll get detailed information about your recommendation model’s quality, top predictive factors, and predicted lift on the scorecard. And once you’re confident about your recommendation’s performance, add business rules in an NBA strategy and deploy your recommendation.

Pretty cool, right? Can’t wait to provide smarter recommendations and more personalized experiences to your customers? Check out this Einstein Recommendation Builder module on Trailhead.

Get started!

The extensive capabilities of the Einstein Platform can be utilized to meet various business needs and use cases. The ones reviewed in this post highlight the ability to create AI and ML models without code, and focus on delivering the predictive insights to all users. From sales executives and call center agents to business analysts and managers, users can take action to improve business outcomes, right where they work. Also, keep in mind that the Einstein Platform is highly scalable, so you can start small and innovate fast as business grows. Einstein is here to help you throughout your organization’s journey.

Finally, check out this trailmix that will help you get started with all things Einstein and its platform features. Enjoy!

Resources

How Salesforce Einstein Is Supercharging Mobile Experiences.

How Salesforce Einstein Is Supercharging Mobile Experiences

While its impact is widespread, one of the most exciting aspects of artificial intelligence (AI) is its ability to create conversational interactions that generate personalized experiences, supercharging productivity and efficiency. In this blog post, we’ll explore how the implementation of large language models on mobile devices is reshaping the enterprise mobile landscape and how Salesforce […]

READ MORE
Einstein standing next to text that says, "How to Use Generative AI Tools to Write SOQL Queries."

How to Use Generative AI Tools to Write SOQL Queries

Salesforce Object Query Language (SOQL) is a powerful tool that allows you to retrieve data from Salesforce. You can use SOQL to query any Salesforce object, including custom objects, custom fields, and user permissions like profile and permission set perms. As a Salesforce Admin, I know that writing SOQL queries can be a pain. Not […]

READ MORE
Headshot of Tom Hoffman next to text that says, "AI Prompt Writing for Salesforce Professionals."

AI Prompt Writing for Salesforce Professionals

The rise of the machines Machines and artificial intelligence (AI) have been part of popular discussion since Samuel Butler authored Erewhon (1872), where his satirical utopian society explored the morality of conscious machines as a natural evolution of the Industrial Revolution. One-hundred and fifty years later, OpenAI released GPT-4, introducing the world to AI that […]

READ MORE