Automated Email Drafting
Lutra enables automated email reply drafting for Gmail by leveraging AI and user-defined instructions. By storing reply guidelines in a separate document (for example, in a Google Doc), you can update the reply guidelines in realtime. A detailed policy document ensures replies are consistent, personalized, and aligned with your communication style.
Common Use Cases
Customer Support: Generate drafts for customer inquiries to help support teams respond faster and more effectively.
Sales Outreach: Create personalized draft responses to leads with relevant product or service information to speed up sales follow-ups.
Internal Communication: Automate draft follow-ups or status updates within teams to streamline internal workflows.
General Queries: Quickly prepare draft replies to common or repetitive questions, saving time for your team.
How to use Lutra
Create a policy document that contains detailed information about how you would like the agent to reply to emails. Example document.
Start a new Chat with Lutra, and ask it to retrieve emails in your Gmail, for example, you can ask it to:
"Retrieve the last 3 emails in my Gmail inbox"
"Retrieve the last 3 emails sent to [email protected]"
Attach (use the 📎 attach button) to attach the policy document to Lutra, and ask Lutra to use the policy document to draft replies.
"Use the attached policy document to draft replies"
Check and refine the behavior of the AI by giving it feedback about the task, or by changing the policy document. If your policy document is provided through a Google Doc or other live data source, ask Lutra to re-read it.
Create a playbook to automate and scale up this process.
Tips for best results
Start Small: It’s always a good idea to test Lutra on a small number of examples before scaling up. Start with a limited dataset or a few emails to ensure the process runs smoothly.
Provide Feedback: Lutra uses AI to adapt and improve based on your feedback. If the initial results don’t meet your expectations, let Lutra know. It can fine-tune its approach, improving both the data it extracts and how it structures the output.
Provide feedback directly, for example:
"The tone of the replies do not match the guidelines, update the writer prompts to be more precise about following this."
Scale and Automate: Once you’re happy with the results, you can scale up the process by asking Lutra to create playbooks to run on many more emails, or on a scheduled basis.
When running on a schedule, have the playbook only search over incoming emails in a time window.
Review Details: You can review the execution logs to see all the details of how Lutra performs tasks. Click on the log entries in the Chat view to expand them to see details.