Automation AppThe Automation Company is your single source for RevOps and Marketing Automation solutions. We empower businesses to achieve more, faster.

Build an agent. Then let that agent build the next one.

8 Apr 2026 · 1 min

Watch on YouTube

0:54

In this short presentation, the speaker highlights Salesforce's recent agent capabilities, emphasizing how they enable precise control over large language model (LLM) outputs—a critical requirement for business use. The speaker demonstrates building an agent and then using that agent to construct a second agent, showcasing a recursive workflow that lowers the barrier for teams unfamiliar with AI development. The talk also promotes the "Ask an Agent" help feature, which simplifies onboarding and encourages users to experiment and test agents without fear of complexity.

Chapters

Agents are all the rage this year, and I was so excited to find out how Salesforce is implementing agents and how I can take that back to my teams to really help them implement as well. Today what we did was build an agent, but what really impressed me was how I could control aspects of the agent. So one of the hardest things with LLMs is that you can't control what they're going to say or what the output is, and in a business setting that is not acceptable. We actually used an agent to build another agent, so it was really cool because even if you didn't know where to start or some of the logic needed, you could just ask AI to even help you build your own AI. If you're scared of building your own agent because you think that it's going to be like way harder or you don't know where to get started, don't be scared. They make it so easy, particularly with the Ask an Agent help. And so like just get started and dive in and test it out.