An Overview of Artificial Intelligence in Salesforce Einstein

By: Collin Couey on December 10, 2019

While far from being a core feature of most CRMs, Gartner found that artificial intelligence (AI) and predictive analytics have the highest interest from businesses across sales, marketing, customer experience, and digital commerce—the four major CRM services (full content available to Gartner clients).

Salesforce is one of the earliest adopters of AI capabilities and technology and the largest customer relationship management (CRM) software vendor, owning 19.5% of the CRM vendor market. In fact, it’s one of the first CRM vendors to have AI capabilities; it launched its AI tool, Salesforce Einstein, in September of 2016 for select modules.

In this piece, we’ll explain what Salesforce Einstein is, as well as provide an overview of the primary features it brings to the table so you can make a more informed decision when purchasing your next CRM.

What is Salesforce Einstein?

Salesforce Einstein integrates AI technology with Salesforce’s Software-as-a-Service (SaaS) CRM. It uses data gathered on every user action to provide predictive analytics, natural language processing (NLP) capabilities, and machine learning to Salesforce customers.

Salesforce Einstein is able to take advantage of its enormous user base by analyzing every action taken to continue to improve its capabilities, which provides users with more accurate analysis as it continues to learn.

For sales, it helps to increase win rate by predicting if a customer is more or less likely to purchase a product with forecasting and lead and opportunity scoring. 

For service, it helps with case classification and routing to increase retention or predict issues. It also includes smart chatbots to help customers resolve common problems.

For marketing, it helps to increase conversion rates by predicting who is more or less likely to engage with an email by engagement scoring and predictive recommendations.

For commerce, both traditional and digital, it uses customer data to advertise what products a visitor might want, helping to increase revenue by predicting if they are more or less likely to purchase a particular item.

Salesforce Einstein has a list of 36 features, some of which are unlocked at different price points, but those features break down into four main categories:

  • Machine learning

  • Natural language processing (NLP)

  • Computer vision

  • Automatic speech recognition

We’ll go into more detail about what each of these include below.

How Salesforce Einstein uses machine learning

Salesforce offers machine learning features in three different tools:

  • Einstein Discovery

  • Einstein Prediction Builder

  • Einstein Next Best Action

These are their predictive analytics tools.

Einstein Discovery derives customer insights using your CRM data as well as other internal and external data sources such as your website, email marketing campaign results, or social media.

It can analyze data more quickly than any manual method in order to help you understand what’s happening with your business and customers, as well as what the next best steps are for your reps.

Einstein Prediction Builder lets users create custom predictions like “What is the likelihood that my customer will churn?” This one is a little more conceptual, so take a look at their short demo video below for more context:

Einstein Next Best Action gives your employees recommendations based on past interactions and collected data. For instance, if Salesforce predicts a customer is likely to quit your services or move on to a competitor, it would tell your employee to contact a particular member of the customer’s team using their preferred method of contact (phone call or email) at their preferred time of day.

If that action seems correct to your employee, Salesforce will then put it on their schedule and set reminders.

How Salesforce Einstein uses NLP

Salesforce classifies its NLP features two ways:

  • Einstein Language

  • Einstein Bots

Einstein Language discovers intent in multiple languages and can detect the sentiment within bodies of text. Salesforce Einstein analyzes the various forms of customer contact with your business such as emails, forms, notes, or chatbot inquiries and scans the text to determine sentiment.

For instance, if it spots the words “happy,” “thrilled,” “excited,” or “overjoyed,” it would determine that your customer is pleased with the product or service. The same is true for negative words, or you might get a score that’s more middling if the text says both negative and positive things.

With NLP, you can find patterns to quickly determine how a customer feels so that you can understand how to approach your interactions with them. If you know they’re upset ahead of time, you will be better equipped to handle the impending interaction.

Einstein Bots work through digital channels and are connected to your CRM. They are smart chatbots designed to improve customer success by answering common questions and increasing your business’s productivity by freeing your employees up to tackle more important tasks.

For more information about what you can do with one of their bots, check out the demo below:

How Salesforce Einstein uses image recognition with computer vision

Salesforce has a variety of image-based AI features called Einstein Vision.

Einstein Vision includes image classification, object detection, and text recognition.

Image classification determines what object is in the image, and object detection determines how many different things are in the image. This feature is particularly useful for businesses that need inventory management or businesses that rent items out to consumers.

For retail businesses, you can take a picture of a shelf of products to quickly determine which objects need to be restocked.

For more information, check out the demo video:

Text recognition can scan an image for text. One practical use for this feature is taking a picture of a business card, which will then upload all relevant contact information from the card into your CRM.

With all three working together, you can train Einstein’s deep learning models to recognize your brand’s logo anywhere online and detect sentiment analysis using text recognition.

How Salesforce Einstein uses automatic speech recognition

The latest in Salesforce Einstein is its automatic speech recognition feature, Einstein Voice.

Einstein Voice connects to Siri, Google, or Alexa and gives daily briefings to your reps. They can get a quick overview of their daily schedule, as well as alerts on any new updates for at-risk customers that they should contact.

You can also update contact information or capture notes with your mobile device on the fly, and Einstein Voice will then give you recommended next steps based on the information you provide. It takes your CRM data and gives your text-to-speech the context of your business.

Salesforce’s demo video provides additional information if you’re curious for more:

Salesforce Einstein cost and availability

Salesforce Einstein is available for all Sales Cloud, Service Cloud, Marketing Cloud, App Cloud, Analytics Cloud, and Community Cloud customers but requires an additional cost to unlock.


All plans are USD per user per month and billed annually

Curious how Salesforce compares to the costs of other CRMs?

Even though Salesforce is a major vendor in the CRM market, there are plenty of other great software options out there to check out. You can start by:

  • Comparing Salesforce with other CRM software options by taking a look at our CRM pricing guide.

  • Looking through our CRM FrontRunners to learn about some of the best CRM vendors out there. (Spoiler alert: Salesforce is one of them.)

If you would rather have a more personalized experience, you can make an appointment with one of our advisors for a 15-minute consultation to see which software would work best for your company.