Prompt engineering for B2B prospecting: structure and examples
· AI & Automation · 7 min
Generative AI can write prospecting messages in a matter of seconds — but the result is only as good as the prompt that drives it. A vague prompt produces a generic message; a structured prompt produces a relevant one. Here’s how to build an effective prompt for B2B prospecting.
Why the prompt makes all the difference
A generative model doesn’t guess your prospect. If you write “write a prospecting email,” you’ll get text that’s interchangeable and the recipient will recognize as such. If you provide the prospect’s industry, job title, a recent update, and the email’s objective, you’ll get a message that looks like a human took the time to write it.
This is the difference between a personalized prospecting email and spam.
The structure of a good prospecting prompt
An effective prompt includes five elements:
1. The role: “You are an experienced B2B salesperson.”
2. The context: who you are, what you sell.
3. The recipient: company, industry, job title, known trigger or recent news.
4. The objective: what to achieve (a conversation, a meeting, a reply).
5. The constraints: tone, length, channel (email or LinkedIn), language, and a ban on jargon.
Example of a structured prompt
Essential guardrails
— Always review and adjust the generated message before sending: AI can invent facts or misinterpret a signal.
— Never let the AI invent a statistic or number: a prospect who checks and discovers a false data point will lose all trust.
— Vary the phrasing from one message to the next to prevent mass sends from looking too similar, which harms deliverability.
Prompt engineering is part of an AI-augmented prospecting workflow (in French) where humans keep the final decision.
Frequently asked questions
- Is a good prompt enough to produce a good message?
- No. The prompt produces a draft. The final quality depends on human review and adjustment, especially to verify the truth of the cited elements.
- Do you need to know how to code to do prompt engineering?
- No. Prompt engineering is a writing task: structuring a clear instruction. No technical skills are required.
- Can AI invent information about the prospect?
- Yes, it’s a real risk. That’s why the prompt must include real information you’ve collected, and the review step must verify every claim before sending.
Prompt engineering is the art of giving AI the context a human would naturally provide. At Mooniz, we structure our prompts using real data about each prospect, then a human reviews every message before sending. Discover our services →