Free template · Outbound & prospecting
The AI-Assisted Outbound Playbook
Use AI for research and first drafts while humans keep control of targeting, claims and sending.
What's inside
- Where AI helps and where a human must stay in charge
- A 7-step workflow from ICP to booked meeting, with owners
- 6 copy-ready prompts for research, personalization and reply handling
- A human review checklist for every AI-drafted message
Who it's for
Founders and first sales hires who want to scale outbound with AI without sounding like a bot.
Why it matters
Most early-stage teams try AI for outbound in one of two ways: they avoid it and spend hours on manual research, or they automate everything and send messages full of invented details. Both fail. The first does not scale past a founder's calendar. The second damages trust with buyers and puts your sending domain at risk.
This playbook draws a clear line between the work AI can take on, such as summarizing a company's site or sorting replies, and the decisions a person must own: who to contact, what to claim and what gets sent. For a small team, that split is what lets outbound grow without losing accuracy or the voice of the company.
How to use it
- 1Start with the task table and agree, as a team, which research, drafting and reply-sorting jobs AI will handle and which decisions stay with a named person.
- 2Run the seven-step workflow in order, writing an owner and a clear 'done when' check for each step so nothing moves forward half finished.
- 3Paste real source text into the research brief and first-line prompts, replace every placeholder in braces, and never ask the model to look up facts on its own.
- 4Put every AI draft through the human review checklist before sending. If a message fails any check, rewrite it or skip that contact.
- 5Review the weekly tracking list, including how many drafts needed a rewrite, and use it to improve your prompts and targeting.
Read alongside it
- How to personalize cold email at scale
- What is a GTM engineer? The role behind modern outbound
- Multichannel outreach: combining email, LinkedIn and calls
The AI-Assisted Outbound Playbook: FAQ
- Can AI write cold emails for B2B sales?
- AI can draft useful first versions, especially opening lines based on real research. It should not send on its own. Models can invent facts about a prospect, and one wrong detail hurts credibility. Treat AI output as a draft, check every fact against a source, and keep a person responsible for what goes out.
- What parts of outbound should never be automated?
- Keep humans on choosing which accounts to target, replying to interested prospects, and anything involving pricing, terms or commitments. Also make sure a person has reviewed every template at least once before it is used. Automation works best for gathering and summarizing information, not for judgment calls or real conversations.
- How do you personalize cold emails with AI without making things up?
- Give the model source material, such as website copy, news or job posts, and tell it to use only that text. Ask it to write 'not found' when the source does not support a point. Then have a person verify each detail and remove anything generic before the email is sent.
More free templates
View all 45Rather have it built than download it?
We build your ICP, CRM and live outbound in a 14-day sprint, documented and handed over.