Automated Outbound Reply Management System

Add contact, company, and phone data to every reply.

• Use Clay to enrich firmographic and contact data.

• Run phone number lookups through BetterContact waterfall (20+ providers).

• Use People Data Labs as a secondary phone source.

• Generate a lead summary with OpenAI for the sales rep.

• Store full enrichment in Supabase for reporting dashboards.

Strip out everything that isn't the actual response.

• Use OpenAI to process the raw reply text.

• Remove HTML formatting and inline styles.

• Remove email signatures, footers, and disclaimers.

• Remove URLs that aren't relevant to the reply.

• Standardize fields across email and LinkedIn replies.

Categorize every reply into one of four categories.

• Use a frontier model (OpenAI) to classify replies as:

  • Interested

  • Neutral

  • Not interested

  • Out of office

• Re-enroll Out of Office replies back into the sending campaign automatically.

• Track classification accuracy and adjust the prompt if the error rate exceeds 5%.

Push sentiment tags back into Instantly and HeyReach.

• Route the data through n8n.

• Look up the correct Slack channel ID using Airtable.

• Update reply tags in Instantly and HeyReach with the classified sentiment.

Notify the right rep with one-click actions.

• Post the reply, lead summary, and sentiment tag to the right Slack channel.

• Add interactive buttons for:

  • Open LinkedIn profile

  • Open company LinkedIn

  • Go to website

  • Open the conversation in Instantly

  • Open the conversation in HeyReach

  • View the CRM record (if using OutboundSync)

  • Mark sentiment as wrong