Use ChatGPT to Write Better Cold Outreach in Minutes

Executive overview

Most BDRs misuse AI by letting it replace their judgment rather than accelerate their process. This episode of The Sales Lab walks through a two-step workflow — research first, then draft — that keeps personalization intact while cutting the time from blank page to send-ready email. The core insight: AI is a first-draft engine, not a finished-email machine. A live demo with a BDR shows how iterative prompting and human editing produce emails that read like they came from a person.

What AI should and should not do

• Use AI to speed up research — surface LinkedIn data, CRM history, and company context faster than manual digging.
• Use it as a brainstorm engine: generate angles, subject lines, and CTAs at volume, then filter with your own judgment.
• Use it to break blank-page paralysis and get to a first draft with momentum.
• Do not let AI replace your thinking — review and edit every output, never copy-paste raw.
• Do not sacrifice personalization — the email must still feel like it came from you after real research.
• Do not let it sound like AI wrote it — generic openers and over-formal phrasing are immediate red flags.

The two-step workflow

• Step 1 — Research the account: Pull prospect and company data from LinkedIn (Sales Navigator if available) and your CRM before touching any prompt.
• Step 2 — Extract two to three real insights: Look for a recent trigger or event, a role-specific pain point, or a stated company priority. Give ChatGPT context, not guesswork.
• Ask ChatGPT to summarize the prospect's role and responsibilities, identify common challenges, and convert raw notes into usable insights.
• Then prompt it to write three to four email variations with different angles, subject lines, and CTAs.
• Also request a 15–30 second cold-call opener with a clear reason for the call.

Live demo: research to outreach

• Starting prompt included the prospect's name, title, company, and LinkedIn page — and asked for three to four email variations plus a talk track.
• First pass was solid on personalization (calling out the prospect's career progression from frontline sales to BDR leadership) but needed tightening on length and the call to action.
• Key edit: replace a weak positive CTA ("would you be open to…") with a negative CTA ("would you be opposed to a 10-minute call this week?") — leverages loss aversion and reads more human.
• Second round of feedback added missing context: who the sender is and what they sell (a sales enablement tool). This grounded the email in a real business reason for reaching out.
• Final output was not perfect — a few sentences still needed trimming — reinforcing that human editing is always the last step.

Key rules to internalize

• AI helps you do your job faster; it does not do your job for you.
• Always give ChatGPT context about yourself, your product, and your goal before asking it to write.
• Iterate with feedback rather than accepting the first output.
• Concise emails with tight CTAs consistently outperform longer, AI-padded ones.
• The standard for any AI-assisted email: does it sound like a human wrote it after actually researching the prospect?