Your First Engine

Creating your first Engine

Now that you're in the Engine Builder and you have named your engine, let's start by adding our first agent to our Engine by clicking the '+' button on the right.

This is where you will define all of the variables for your new agent. Give it a name, choose which LLM model it will leverage, insert your prompt and choose the tools it will use.

It is a best practice to not overload one agent with too many different tasks, to achieve the best result it is best to segregate the agents by task and allow each one to have a singular function.

In our example, we are creating an engine to do cold email outreach, therefore our first agent will research new leads which will later be passed to an email writer agent (if you already have an existing list of leads, upload them into your Google drive and use the 'Google Sheets' Tool to query them).

After adding your agent, it is best to test the output of your engine, to ensure your agent is producing the expected result. The next step is where the magic happens.

Next, we will create a 'Client Outreach' agent. This agent will be responsible for crafting a personalized email and drafting it in your Gmail mailbox (you can use the scheduler as well if you want the agent to automatically send your email).

We now prompt the chat to test out our engine, looks like we were successful in finding a lead, let's see how the email draft looks.