The problem OutreachAI solves
The Problem It Solves
- Reaching the right people: whether for podcasts, partnerships, or sales, the manual process is slow, and inefficient.
- You waste time searching: Finding relevant people across the internet is a scattered mess.
- Cold outreach feels like a shot in the dark. Messages can go ignored if they lack personalization or context.
- Follow-ups and scheduling are tedious. Even when there's interest, managing back-and-forth can kill time.
- Outreach at scale needs teams. Doing this for multiple prospects or campaigns usually demands resources and lots of manual effort.
What OutreachAI Makes Easier
- Autonomous Discovery: AI agents search across platforms to find the right people based on your intent.
- Personalised Messaging: Each mail message is tailored to the recipient’s background, increasing the chances of a reply.
- Follow-up & Scheduling: It handles reminders, reschedules, and meeting setup so you don't have to.
- Crypto Payments: Seal the deal with autonomous crypto payments directly from agent's CDP wallet which you fund while creating campaign.
Challenges we ran into
Finding contacts across platforms
- Challenge: Prospects exist across linkedIn, twitter, youtube, personal websites, etc. Finding them is a challenge, especially their emails.
- How We Solved It: We partially tried to solve this by making firecrawl specifically search the about/profile pages, where mostly contact mails are present.
AI Agent Debugging
- Challenge: When something went wrong (e.g. no replies, duplicate messages, hallucination), debugging autonomous agent behavior was painful.
- How We Solved It: We added shadow logs could show agent decisions step-by-step, and which tools were used, etc. This gave us clarity.
Mail replies
- Challenge: Whenever a user replied to the mail sent by agent, follow-back is tricky to implement due to the asynchronous and external nature of email.
- How We Solved It: We used Google Cloud Pub/Sub to subscribe to Gmail webhook notifications via Gmail API.