The Prompt Recipes Every SDR Should Have Saved

Prompting an AI model is a skill. The difference between a useful output and a useless one isn't which tool you use — it's how specifically you describe what you want.

Most SDRs treat AI like a search engine. They type a question or a vague instruction and hope for something usable. That works occasionally. But the teams who get consistent, high-quality outputs have learned to treat every prompt like a brief — with an audience, a goal, constraints, and a format.

This article is a collection of 7 specific prompt recipes for the tasks SDRs do most — built with the structure that actually produces usable results.

What Makes a Good Prompt

Before the recipes, here's the structure behind all of them. Every effective prompt has four elements:

1. Audience — Who is this for? The more specific the role, company type, and context, the better.
2. Goal — What should this output accomplish? A reply, a list, a summary, a reframe?
3. Constraints — Word count, tone, what to avoid. Constraints improve output quality more than anything else.
4. Format — Bullet points, numbered list, paragraph, table? Tell the AI how to present the output.

Now the recipes.

Recipe 1: Build a Prospect List

Use this when you need to define filtering criteria for a new outreach campaign.

Why it works: It forces specificity before you start pulling leads — which improves targeting and reduces wasted outreach on low-fit accounts.

Follow-up prompt: "What's a buying signal I should look for in this ICP that would tell me they're actively evaluating solutions right now?"

Recipe 2: Write a Cold Email Sequence

Use this when building outreach for a new segment or campaign.

Why it works: The constraints (word count, tone, varied CTAs) prevent the common AI failure mode of producing four emails that all sound like the same email.

Follow-up prompt: "Rewrite Email 2 to focus on a specific outcome rather than features. Use a real-sounding metric like '60% less time' rather than vague claims."

Recipe 3: Generate Subject Lines

Use this when open rates are low or you need variety to A/B test.

Why it works: Asking for mixed styles in one prompt gives you testable variety without multiple separate requests.

Follow-up prompt: "Which of these would most likely get opened by someone who gets 100+ cold emails per week? Rank the top 3 and explain why."

Recipe 4: Handle Objections

Use this when a prospect replies with a common pushback and you need a strong response fast.

Why it works: Three approaches gives you options to choose from based on tone, relationship, and what you know about the prospect.

Follow-up prompt: "Take the reframing version and make it shorter. The reply should feel like it came from a confident rep who's heard this before, not someone scrambling to justify."

Recipe 5: Write Follow-Up Emails (No Reply)

Use this when a prospect has gone quiet and standard follow-ups aren't getting responses.

Why it works: Explicitly telling AI not to reference non-replies avoids passive-aggressive language that kills follow-up open rates.

Follow-up prompt: "Give me a version that leads with a question instead of a statement — something that makes them want to answer even if they weren't interested before."

Recipe 6: Summarize Calls and Deals

Use this after a discovery call to create a structured summary for CRM notes or handoffs.

Why it works: The 4-section structure ensures nothing operationally important gets buried in a wall of text, and the bullet limit keeps it scannable.

Follow-up prompt: "Based on this summary, what's the most important thing I should address in my follow-up email to move this deal forward?"

Recipe 7: Analyze Campaign Performance

Use this after a campaign completes to extract actionable insights from the metrics.

Why it works: Asking for "2–3 reasons" prevents the generic "here are 10 things to improve" list. Asking for one specific test per issue forces actionable output.

Follow-up prompt: "If I could only fix one thing in the next campaign, what would you prioritise based on the data? Why?"

The One Rule That Applies to Everything

Add a constraint to every prompt. Word limits, tone requirements, format rules — they all dramatically improve what AI returns.

Without constraints, AI defaults to the longest, most general, most hedged version of whatever you asked for. With constraints, it's forced to make choices — and those choices almost always improve quality.

If you're getting mediocre output from any of these recipes, the first thing to check isn't whether the prompt is well-written — it's whether you've added enough constraints. Tighten the word limit. Specify the tone. Tell it what not to include. The output will improve.

Save these recipes. Adjust them for your ICP and product. Use them every day. The SDRs who get the most out of AI aren't the ones who know the most about AI — they're the ones who've built a reliable set of prompts that consistently produce usable output.