Step 1: Create a New Agent
To get started with your company research agent:- Navigate to the Agent Builder section in Agent.ai
- Click “Create Agent”
- Select “Start from scratch”
- Name your agent (e.g., “Simple Company Research”)
- Add a description like “This agent will take the name of a company and do some research on what products the company sells”
Step 2: Set Up the Trigger
For this simple agent, you’ll use a manual trigger:- In the trigger section, select “Manual”
- This means users will manually start the agent when they need company information
Step 3: Create the Input Action
Now you’ll set up how the agent collects the company name:- Go to the Actions screen
- Click to add a new action
- Select “Inputs & Data Retrieval”
- Choose “Text box” as the input type
- Add a prompt like “Enter the name of the company”
- Optionally, add an example (e.g., “Upsot”)
- In the “Store value in” field, enter “out_company” (this creates a variable to store the company name)
- Save this action
Step 4: Create the Research Action
Next, set up the AI to research the company:- Add another action
- Select “AI & Content Generation” then “Generate Content”
- Choose an LLM model (e.g., GPT-4 Mini for cost efficiency)
- In the instructions field, enter your prompt:
- You are an experienced researcher. I am giving you the name of a company: out_company. I would like you to research this company and list out the products this company sells and what industry it is in.
- In the “Store output in” field, enter “out_company_research”
- Save this action
Step 5: Display the Results
Finally, set up how the agent will show the research results:- Add a final action
- Select “Outputs”
- Choose the variable “out_company_research” to display
- Save this action
Step 6: Test Your Agent
Your company research agent is now ready to use:- Click the “Run” button to test your agent
- When prompted, enter a company name (e.g., “Pigment.com”)
- The agent will process the request and display information about the company’s products and industry
Example Output
When testing with “Pigment.com”, the agent might return something like: “Pigment operates in the field of business intelligence and data visualization, focusing on helping organizations manage their data more effectively. Their main product is a SaaS platform that offers business planning, forecasting, and analytics solutions.”Advanced Options
Once you’re comfortable with your basic agent, you can explore more advanced features:- Set up automatic email generation of results
- Create more complex research workflows
- Add additional data sources
- Customize the output format