A vague prompt gets you a long list. A precise one gets you a short list of people worth talking to this week.
AI prospecting tools have made it possible to describe your ideal customer in a sentence instead of a stack of filters. That’s great, until you realize the quality of the list now depends on the quality of the sentence.
We’ve watched a lot of briefs go through Manet. The ones that come back with great people share a pattern. They’re specific about who, they say why now, they leave out anything nobody could verify, and they say who to exclude. Below is that pattern, followed by 25 prompts you can copy and adapt. They’re written for Manet, but most will work in any tool that accepts natural-language search.
The formula behind a good prospecting prompt

Who is the role and seniority. Where is the type, size and location of the company. Why now is a signal with a time window, like a post, a hire, a funding round or a job change. Not is who to leave out.
The test for every part is simple: could something public confirm it? “Hiring SDRs” can be checked against job listings. “Has budget” can’t. “Bootstrapped” can’t be proven directly, but “hasn’t raised outside funding” can be checked by looking for funding rounds. If a requirement can’t be confirmed, a good tool will either ignore it or come back with nobody. We wrote about why in how Manet’s people search works.

Prompts for role and company fit
Start here when your ICP is clear and you mainly need the right people at the right companies.
“Find heads of RevOps at B2B SaaS companies with 50 to 300 employees in the UK and Ireland.” A clean baseline. Add a signal once you see who comes back.
“Find founders of design or development agencies with 10 to 50 people in Western Europe.” Good for service businesses where the founder is still the buyer.
“Find VPs of Sales at fintech companies in London that sell to banks or insurers.” “Sell to” is a strong filter that most title searches can’t express.
“Find CFOs at Series A to Series C healthtech companies in the US.” Funding stage is checkable and narrows company maturity quickly.
“Find heads of people at companies with 200 to 1,000 employees that are fully remote.” Remote status is often stated on the careers page, so it can be verified.
Prompts for timing signals
These find people at the moment they’re most likely to care. If you only use one section, use this one.
“Find sales leaders at B2B software companies that have posted SDR or BDR job openings in the last 30 days.” Hiring for a role is one of the clearest signs of a priority. More in how to find companies hiring SDRs.
“Find founders of companies that raised a seed or Series A round in the last 90 days in France or Germany.” New money usually means new hires and new tools. Keep the window recent.
“Find people who started a new role as head of marketing in the last 60 days at B2B companies.” New leaders review their stack and their agencies in the first few months.
“Find VPs of Engineering at companies that announced an office or market expansion this quarter.” Expansion creates projects, and projects need help.
“Find customer success leaders at SaaS companies hiring more than three CSMs right now.” Volume of hiring tells you how big the push is.
Prompts for posts and engagement
People who post or comment about a problem have told you, publicly, that they care about it. Use that, gently. See how to find people who recently posted about your topic for the manual version.
“Find heads of sales who posted about outbound reply rates or cold email in the last 30 days.” The post becomes the reason for your first comment.
“Find founders who commented on posts about founder-led sales in the last two weeks.” Commenters are often less visible than posters, and less pitched to.
“Find people who engaged with my latest post and work in sales at a B2B company.” The warmest possible starting point. Lists like this tend to come back fast because the starting pool is small.
“Find people who commented on [link to a specific post] and are directors or above.” Great for a big industry post or a competitor’s announcement.
“Find marketing leaders who regularly post about attribution or pipeline reporting.” “Regularly” separates practitioners from one-off posters.
Prompts for network and warm paths
Your existing network is usually the most underused source of pipeline.
“Find second-degree connections who are heads of growth at Series A or B startups.” Second-degree means a shared connection you can mention or ask.
“Find people in my network who changed jobs in the last 90 days and now work at companies with more than 100 employees.” Old relationships, new budgets.
“Find former colleagues from [company] who are now in a sales or revenue leadership role.” People you’ve worked with reply at a much higher rate.
“Find first-degree connections I haven’t messaged in over a year who fit my ICP.” Reconnecting is easier than cold outreach, and nobody does it.
Prompts for events and community
“Find heads of sales in Berlin who posted about attending SaaStr Europa or similar sales events this year.” Useful before and after an event you’re going to.
“Find founders of B2B startups within 50 km of Lisbon to invite to a dinner on outbound.” Exactly the kind of brief on our homepage: find, then invite.
“Find people who spoke at or organized sales or RevOps meetups in London in the last 12 months.” Speakers and organizers are connectors. One good relationship opens many doors.
Prompts for refining a list you already have
Most of the value comes from the second and third version of a brief, not the first.
“Same brief, but exclude agencies, consultancies and recruiters.” The fastest way to clean up almost any list.
“Same brief, but only companies that sell to other businesses, and exclude current customers in my CRM.” Exclusions are underrated. Ask for them explicitly.
“Same brief, but only people whose current role started more than six months ago.” Sometimes you want the opposite of a job-change signal: someone settled enough to buy.
Prompts that don’t work, and how to fix them
Instead of | Try | Why |
|---|---|---|
“Founders under 35” | “Founders who started their company in the last three years” | Age isn’t public. Company age often is. |
“French founders” | “Founders of companies based in France” | Nationality can’t be verified. Company location can. |
“Bootstrapped companies” | “Companies with no outside funding rounds on record” | You can’t prove a negative, but you can check for its opposite. |
“Companies with budget for sales tools” | “Companies hiring two or more salespeople right now” | Budget is private. Hiring is public. |
“People who would love our product” | “People who posted about [the problem] in the last 30 days” | Love isn’t checkable. Talking about the problem is. |

Five tips for writing your own
Write it the way you’d brief a colleague. Natural language is the point. You don’t need keywords or syntax.
One signal at a time. Stacking four signals often leaves nobody. Add them one by one and watch the list.
Always include a time window. “Recently” means different things to different tools. “In the last 30 days” doesn’t.
Ask for the evidence. A good result tells you why each person matched. If it can’t, treat the match with suspicion.
Iterate quickly. If your tool takes ten minutes per search, you’ll settle for the first list. If it takes under a minute, you’ll find the fourth version is the good one.
Every prompt here works better on top of a clear ICP. If you don’t have one yet, use our free ideal customer profile template.
Frequently asked questions
What is an AI prospecting prompt?
A plain-language description of the people you want to find, which an AI prospecting tool turns into searches and checks. A good one covers who, where, why now, and who to exclude.
What's the best prompt for finding B2B leads?
One that combines a clear role and company with a timing signal and a time window, for example: "Find sales leaders at B2B SaaS companies with 50 to 300 people that posted SDR job openings in the last 30 days."
Can I use these prompts in ChatGPT?
General chat assistants can help you brainstorm an ICP or draft messages, but they can't reliably search LinkedIn or confirm who currently works where. Use a prospecting tool that checks requirements against live evidence.
Why does my AI prospecting search return nobody?
Usually because one requirement can't be confirmed from public information, such as age, nationality or budget, or because too many signals are stacked at once. Remove or rephrase one requirement at a time.
How many requirements should a prospecting prompt have?
Three to five is a good range: a role, a company description, one signal with a time window, and one or two exclusions.
Do these prompts work in Manet?
Yes. They were written with Manet in mind. Manet turns each part into a requirement, checks people against evidence, and tells you if a requirement can't be confirmed.
Pick one prompt, swap in your ICP, and run it in Manet. Then change one line and run it again.
[
ready to start?
]






