AI for Outreach and Prospecting

Why most AI-generated outreach fails

The promise of AI-generated outreach is personalization at scale. The reality, for most teams, is generic messages that feel like templates because the AI had no real information about the recipient. "I noticed your company is growing fast" is not personalization. It is the illusion of it.

AI generates good outreach when you give it real, specific information. The input quality determines the output quality. Garbage in, garbage out is never more true than in sales emails.

The right input for outreach generation

For each prospect, gather at minimum: their recent LinkedIn activity or posts, one piece of recent company news, and the specific business problem you believe they have. Then prompt: "Write a 3-sentence cold email to [name] at [company]. Context: [what you know about them]. The email should reference [specific thing] and connect it to [the problem we solve]. End with a low-friction ask — a 15-minute call, not a demo. No subject line yet."

Then ask: "Now write 3 subject line options. Short, specific, no hype."

Follow-up sequences

Most sales require 5-8 touches before a response. AI makes it easy to write a full sequence at once: "Write a 5-email follow-up sequence for someone who received the first email but did not respond. Each email should: add new value (don't just follow up for the sake of it), be shorter than the previous, and shift the framing slightly. Emails 3 and 4 should approach from a different angle than emails 1 and 2."

A/B testing your messages

Give ChatGPT your best-performing message and ask: "Generate 3 variants that test different opening hooks, while keeping the core value proposition the same." Run the variants in your sequencing tool and let performance data tell you what's working. Iterate from results, not from opinions.