Artificial Intelligence / Admins / Consultants / Developers

Everything You Need to Know About Salesforce Einstein Model Builder

By Christine Marshall

Einstein Model Builder is a feature of the Einstein 1 Studio that enables you to connect your data with your preferred large language model. 

In this article, I’ll explain what large language models are, what they mean in the context of Salesforce, and how to get started with Einstein Model Builder. 

What Are Large Language Models?

Large language models (LLMs) are advanced artificial intelligence systems that use deep learning techniques to understand and generate human-like text. LLMs are trained on vast amounts of data to learn the nuances of human language, including grammar, syntax, semantics, and context. Although they are super smart, LLMs do not understand as humans do – they remember data and connect the information to present you with an answer. 

The term “large” refers to the vast scale and complexity of the model. These models typically have hundreds of millions to billions of parameters, allowing them to capture and understand complex patterns in language and return better-quality responses.

  • Training: LLMs are trained to become more effective. This involves feeding more information into the LLM, asking it to make predictions, testing it, and giving feedback on answers so it learns what is correct and what isn’t.
  • Fine-tuning: LLMs can also be “fine-tuned” to become experts in a certain subject. Initially, the LLM may be fed a large dataset to learn from. Fine-tuning involves feeding the LLM more specific or unique information on a particular subject.
  • Versions: You’ll also notice that LLMs typically have a version number. This is because providers of LLMs are continuously improving or refining the LLM. Each version benefits from the learnings in previous versions but becomes more sophisticated and refined over time. 

Large language models can perform various natural language processing tasks, such as text generation, translation, summarization, question answering, sentiment analysis, and more. They are used in a wide variety of applications, including virtual assistants, content generation, language translation, and automated customer support.

In the context of Salesforce, large language models can be applied in several ways to enhance the capabilities of the Salesforce platform and improve user experiences. Here are some examples:

  • Personalized Content: LLMs can be used to generate personalized content, such as emails to customers. LLMs can also create product recommendations based on customer data, such as purchase history or cases created. 
  • Automating Tasks: LLMs can be used to automatically categorize and route customer queries, extract relevant information, and suggest appropriate responses or actions for customer service representatives.
  • Text Analytics: LLMs can analyze large amounts of text data stored within Salesforce, such as customer feedback, survey responses, or support tickets. Using this information, LLMs can uncover insights and help businesses make data-driven decisions.
  • Automated Documentation: LLMs can assist in automating the creation and maintenance of documentation, knowledge bases, and internal resources within Salesforce. For example, an LLM could be used to create Salesforce Knowledge articles based on Cases. 

What Is Einstein Model Builder?

Salesforce Einstein Model Builder is a component of Salesforce Einstein 1 Studio, which is an artificial intelligence (AI) platform within Salesforce’s core platform. Model Builder allows Salesforce professionals to create custom AI models tailored to specific business needs without requiring extensive data science expertise or integrations.

With Model Builder, you can leverage your Salesforce data to train machine learning models for various purposes. The platform provides a user-friendly interface and guided workflows to facilitate the model-building process.

Salesforce follows an agnostic approach to large language models (LLMs). You can use Salesforce’s Shared LLMs, Salesforce’s hosted third-party LLMs, or bring your own model (BYOM) from other platforms (e.g. Amazon SageMaker, Google Vertex AI) and combine your company data with preferred AI models from those other predictive or generative AI services.

Top Tip: Different use cases or data residency concerns may require different LLMs. That’s why it is so important that Model Builder lets you choose which LLM to use! 

  • Salesforce Shared LLM: Salesforce uses a secure gateway to travel across the internet to ask its LLMs for a response, including CodeGen, CodeT5+, and CodeTF.
  • Salesforce Hosted Third-Party LLM: You can host LLMs from Amazon, Anthropic, Cohere, and others within the Salesforce infrastructure.
  • Bring Your Own Model: Connect your own model to Einstein through the Einstein Trust Layer. 

In essence, Einstein Model Builder enables you to combine your company data with your preferred LLM from other predictive or generative AI services, giving you complete control. You manage your data and LLMs from a unified control plane using clicks, not code. 

Einstein Model Builder In Action

There are a few steps to create a model using Einstein Model Builder:

  • Access
  • Build
  • Utilize
Source: Use AI Models
  1. Model Builder will prompt you to choose the type of model.
  1. Select the training data and define any filters.
  1. Set the goal for your model.
  1. Prepare the data by selecting which fields to include in the model as variables.
  1. Choose your algorithm.
  1. Review and train your model.
  1. Click Save & Train to create the model.

Once the model is activated, you can see the details on its record page. This includes prediction jobs that keep a log of the model’s usage within your CRM. Here you will see useful information including:

  • Job Name
  • Type
  • Status
  • Last Run Date
  • Last Run Status



Einstein Model Builder enables you to combine your company data with your preferred LLMs in a safe, unified platform. You have complete control as the customer, and can easily manage your models using clicks, not code.

The Author

Christine Marshall

Christine is the Courses Director at Salesforce Ben. She is an 11x certified Salesforce MVP and leads the Bristol Admin User Group.

Leave a Reply