Step 1: Create a New Agent
- Start by creating a new agent in the Builder Tool.
- Name your agent “Executive Tweets”
- Add a description like “Given the Twitter handle of an executive, this agent will return a summary of their tweets”
- Set the trigger as “Manual”
Step 2: Get the Twitter Handle
- Add a “Get User Input” action
- Set the prompt to “Enter the handle of the executive”
- Add examples like “@username” to guide users
- Set the output variable name to “out_handle”
- Save this step
Step 3: Fetch Recent Tweets
- Add a “Social Media & Online Presence” action
- Select “Recent Tweets” from the available data sources
- Set it to fetch the 20 most recent tweets
- Insert the handle variable you collected in the previous step
- Set the output variable name to “out_tweets”
- Save this step
Step 4: Generate a Summary
- Add a “Generate content” action (you can use Claude or another available model)
- Create a prompt like: “I am providing you with a list of tweets. Please generate a summary about these tweets.”
- Pass in the tweets variable from the previous step
- Set the output variable name to “out_summary”
- Save this step
Step 5: Display the Results
- Add a “Show User Output” action
- Include both the summary and the raw tweets for reference
- To improve readability, consider adding formatting like line breaks between tweets
- Save this step
Step 6: Test Your Agent
- Run your agent
- Enter an executive’s Twitter handle (e.g., @username)
- Review the summary and raw tweets that are returned
- Main topics the executive discusses
- Their engagement patterns
- Any recurring themes or interests
- Recent announcements or news they’ve shared
Tips for Improvement
For better formatting and readability, you can:- Add line breaks between tweets in your output
- Use the LLM to format the tweets in a more structured way
- Create a more detailed prompt that asks for specific insights about the executive’s Twitter activity