Prompt like an operator

Overview

Filters can only sort by details already saved as fields, like industry, size, or location. When the account you want is defined by meaning, a mix of traits, or something that just happened, filters run out of room. Prompting gets past that: you describe the audience in plain language and the agent builds the search. Your list quality comes down to how clearly you ask.

What you can ask for that no filter can

When you describe an audience in words, the agent builds that search on the spot, even for things no dropdown covers:

• People by what they have done, like a first sales hire at an earlier company.
• A count and a cutoff you set, like more than five revenue operations hires who all started this year.
• The shape of a team, like a high ratio of account executives to sales development reps.
• What is changing, like hiring for a role speeding up.
• Companies similar to your best customers.

Anatomy of a strong prompt

A clear request reads like a short brief. Name four things:

• Who the people are: the role, seniority, or function.
• What kind of company: industry, size, location, or technology.
• What recent event matters: funding, a hiring jump, a leadership change.
• What to exclude: current customers, or accounts already in your CRM.

VP of Sales at US B2B SaaS companies that raised funding in the last 6 months, excluding current customers.

Start wide, then narrow

The most common mistake is trying to describe the perfect list in one prompt. A tight one-shot often comes back thin, because the narrower the first pass, the more good accounts it quietly leaves out. Work the other way, the way a go-to-market team sizes a market:

• Start wide. Describe the whole addressable universe, for example every company in a category. Aim for coverage, not precision.
• Narrow with AI qualification. Ask a question the agent researches per account, like is this a SaaS business. This keeps good-fit accounts a plain industry filter would miss, and drops the ones that only looked right.
• Score and prioritize. Ask the agent to rank the qualified list so your team works the best accounts first.

Three habits keep it reliable:

• Check the sample first, before the agent builds the full table.
• Test qualification on 10 to 20 accounts before running the whole list.
• For a niche audience, seed a few known-good companies and ask for lookalikes.

Refine as you go

The first result is a starting point. Adjust in the same conversation instead of starting over, since each reply remembers the last. Give the agent your ideal customer profile or your best customers up front, and every result improves.

Common mistakes

• Vague requests return vague lists, so name the role, company type, and event.
• Trying to nail the whole list in one prompt leaves out good accounts, so start wide and narrow in stages.
• Leaving out what to exclude brings back accounts you already have.
• Describing the tool instead of the result limits the agent, so say what you want, not how to get it.

Prompts to try

Heads of demand generation at US B2B SaaS companies with 200 to 1000 employees that raised a Series B in the past six months.

Founding sales hires at B2B SaaS companies, referencing first sales hire or zero to one in any past role. Exclude founding engineers.

Companies where hiring for AI and ML titles is more than double the prior 30 days.