At TDX 2025, there were some amazing speaker sessions focused on improving the developer experience. These were my highlights of the latest and greatest in Agentforce:
1. My favourite aspect of Agentforce 2dx is the ability to proactively have conversations with customers or employees without requiring them to initiate the conversation. Agentforce can be triggered by real-time data updates. For example, a company can connect their IoT product’s data to Agentforce with Data Cloud. If there are any issues with connectivity of the product, Agentforce can be leveraged to send out communication to a customer proactively to ask them to restart their internet.
2. Another great feature, stemming from bringing Google’s Gemini product to Agentforce, is the multi-modal capabilities allowing Agentforce to consume text, images, audio and video.
3. They announced Agentforce Developer Edition, a 10GB Developer edition for Data Cloud and Agentforce, which also includes 150 Large Language Model generations per hour. LLM actions include:
- Responding to a user query
- Automated action based on business logic
- A summary or a recommendation
These updates are exciting because they allow developers to experiment with LLMs in their sandboxes using a larger set of data than previously possible, all while also being able to harness the power of Data Cloud. The only catch is that you have to make sure to log in to the org every 45 days (or set up an agent to do that for you).
4. The ability to leverage Agentforce with an External Service in Flows is extremely valuable. Overall, I think Salesforce hit it out of the park with the hands-on sessions and minihacks at TDX 2025. One of my favourite parts was the minihack that required an HTTP callout to an external service for providing estimates on a car trade-in.
This minihack was great, as it had three levels of complexity, building on the previous level’s completion. The third level required setting up an Experience Cloud site, where a customer could interact with the AI agent and ask for a trade in value for their car. The agent was deployed to the Experience Cloud Site through the Embedded Services Deployment.
5. While not exactly released at TDX 2025, it was great to visit the Slack booths and learn all about the impressive AI capabilities Slack offers. My favourite feature by far is ‘Recap.’ It summarizes Slack conversations and threads for the user in a simplified, actionable way. And while the recap function is great for coming back from vacation and getting a summary of what you’ve missed, it’s just as great for day-to-day. For those in meetings all day and have hundreds of Slack messages to catch up on, it’s a great way to focus on the action items. Sometimes I go through a conversation of close to 50 messages back and forth between team members about an issue that came up, and then realize everything has been resolved and does not need my input. With the recap function, I can know this immediately and focus on threads or issues that need my attention. As someone who is included in MANY Slack channels, this feature is a must.
These TDX 2025 sessions demonstrated Salesforce’s commitment to enhancing developer capabilities through AI integration. The advancements in Agentforce are incredibly valuable for Salesforce users across the organization, and represent a major step forward in empowering developers to create more sophisticated, AI-driven solutions while continuing to focus on practical business applications.
Wondering how Agentforce can help your organization? Get in touch. We’d love to connect.
