Salesforce Spring ’25 Release is here, bringing exciting top features for Salesforce developers! This is Part 1 of our two-part series, where we’ll explore the most important updates and how they can benefit you. We’ve got you covered whether it’s smarter AI, better integrations, or enhanced security.

Salesforce Spring ’25 release is rolling out to sandboxes and production environments from January to February 2025. This update brings exciting new features to enhance your development experience.

Let’s get started and prepare for what’s ahead!

Agentforce: Build Smarter AI-Powered Agents

Agentforce makes it easy to create and customize AI-driven agents that work across Salesforce, Slack, websites, and third-party platforms. The Salesforce Spring 25 release adds powerful new features:

  • Agentforce in Scratch Orgs – Now, you can develop AI-powered agents in scratch orgs using Salesforce’s generative AI tools, including Agentforce, Prompt Builder, Model Builder, and the Models API. Just set up a generative AI scratch org, assign the necessary permission sets (like Prompt Template Manager for Prompt Builder), and start building.
  • Agentforce Testing Center – A brand-new tool to test your agents in bulk! You can auto-generate test cases using AI or upload your own. Future updates will bring CLI support and APIs for even more flexibility.

spring 25 - 1

Preview Conditions in Agent Builder:

Now, you can fine-tune your agent testing in Agent Builder by specifying the context of a conversation. This helps you test how Agentforce (Default) or an Agentforce Service Agent (ASA) responds in different scenarios.

  • When testing Agentforce, you can simulate conversations based on factors like the user’s language, app, or the current page type—even down to a specific object or record.
  • Agentforce (Default) is simply the new name for Einstein Copilot for Salesforce.

For ASA, you can test interactions based on a specific end user’s ID or contact ID. If your ASA is connected to Messaging for In-App and Web channels and uses context variables, you can also simulate conversations using session variables you’ve added to your agent.

This makes testing more precise, so your AI-powered agents perform exactly as expected in real-world scenarios.

spring 25 - 2

LLM Integration Recipes

The Salesforce LLM Open Connector lets developers connect any LLM to Salesforce using an Open API specification, making it easier to build custom actions for Agentforce. To help you get started, we’ve added step-by-step process that show you how to bring your own model from your preferred inference provider.

Agentforce API

Exciting news—Agent APIs are launching in February! These APIs let you invoke Agentforce headlessly from other agents and applications, opening up new possibilities for automation and integration. Want a sneak peek? Watch this video for a preview of what’s coming.

Agent Versioning & Management

Managing your agents just got easier! Now, you can create up to 20 versions of your agent for testing, troubleshooting, or saving backups—without affecting the active version. This ensures a smooth iteration process and quick recovery if needed. This feature will be available in the February monthly release.

Data Cloud Updates:

Seamless JDBC Connectivity for Data Cloud

The Java Database Connectivity (JDBC) driver now lets you connect directly to Data Cloud’s ANSI SQL API, so you can access and retrieve data using your favorite JDBC-supported tools.

Enhanced Data Cloud REST APIs

We’ve added new Data Cloud Connect REST APIs to make it easier to create activations and validate transforms. These updates help streamline workflows and improve efficiency when working with Data Cloud. Here are some of the new APIs now available:

spring 25 - 3

You can learn more about the Data Cloud REST APIs in the docs, or explore them in our updated Postman collection.

Enhanced Data Cloud Flow Capabilities:

Flow Builder now supports more Data Cloud data types, including Email, URL, Phone, Percent, Boolean, and Currency. You can use Data Model Object (DMO) and Calculated Insight Object (CIO) fields with these new data types to:

  • Set entry conditions in Data Cloud-triggered flows
  • Send data to Data Cloud with the Send to Data Cloud action
  • Specify filter conditions in the Get Records element

Plus, you can now deploy changes made to Data Cloud-triggered flows from sandbox to production, making testing and deployment much smoother.

Data Cloud One: Now with Enrichments:

In the last release, Salesforce introduced Data Cloud One, allowing multiple orgs to share a unified customer profile and Data Cloud-powered features through a single centralized Data Cloud. Now, with the latest update, you can add enrichments—which means you can pull valuable Data Cloud insights directly into Salesforce CRM record pages using related lists or a copy fields feature. This enhancement will be available in February.

Leverage Data Cloud for Smarter Prompting:

Now, Prompt Builder lets you ground prompts using Data Cloud data graphs and Data Model Objects (DMOs). Data graphs bring information from multiple sources into a unified dataset, giving you a 360-degree view of your customer data. With this deeper integration, you can create more precise and insightful AI-powered prompts in Prompt Builder.

Here’s a preview of how you can ground prompt templates with data graphs in Prompt Builder 👇

spring 25 - 4

Wrap-up

We’re just getting started with the Salesforce Spring ’25 Release! In this first part, we explored key features that can enhance your development process. But there’s more to uncover! Stay tuned for Part 2 we’ll reveal even more exciting updates and expert insights.

Get ready to level up your Salesforce skills with The Pinq Clouds available on AppExchange!