THIS ARTICLE IS FOR: ✅ Self-Serve
Stage: Trial / Onboarding / Live
Owner: CS
Last updated: 2025-12-15
TL;DR
AI Prompting lets you build targeted company lists using one short written prompt.
The quality of your leads depends on how you write the prompt.
Keep prompts short, specific, and written in your market’s language.
You can skip manual filters and let AI do the heavy lifting.
When you’d use this / Why it matters
Use AI Prompting when you want to quickly generate high-quality B2B company lists without manually stacking filters. Well-written prompts unlock larger TAMs, better relevance, and faster list creation.
What is AI Prompting?
AI Prompting allows you to describe your ideal company in plain English.
Example prompt:
“Agencies helping roofers book more appointments”
ListKit’s AI then:
Interprets what those companies do
Matches real-world company descriptions
Returns a targeted list—no filters required
This replaces the need for Keywords and Industries filters.
Step-by-step: Where to use AI Prompting
Go to B2B Search
Click Companies
Open Filters
Select Context → Description Search
Paste your prompt and run the search
B2B Search → Companies → Filters → Context → Description Search
The goal of AI Prompting
Your prompt tells the AI:
What type of company this is
Who they serve
What problem they solve
The clearer that picture is, the better the results.
Example:
❌ “Marketing companies”
✅ “Marketing agencies helping dentists get more patients”
Use our Custom GPT to generate perfect prompts (recommended)
You don’t need to brainstorm prompts manually.
ListKit built a Custom GPT Prompt Builder that:
Takes a company website URL
Instantly generates 5 optimized prompts
Uses patterns proven across thousands of searches
What the GPT optimizes for:
ICP accuracy
Broad but relevant TAM
Plug-and-play prompts for ListKit
Example input:
Calendly.com
Example outputs:
“Software that helps businesses schedule meetings automatically”
“AI tools that streamline team scheduling”
“Platforms for automating client appointments”
Paste any of these directly into Description Search.
👉 Use the Custom GPT here: ListKit Prompt Builder
📸 Screenshot to add here:
Custom GPT interface + example outputs
Core principles of a perfect prompt
1. Be specific (niche beats broad)
Specific prompts = better list quality.
✅ “Software companies selling to dentists”
❌ “Software companies in the US”
Why this works:
Higher relevance
Easier personalization
Less wasted outreach
2. Keep it short (6–9 words max)
Long prompts reduce accuracy.
✅ “Software helping eCommerce brands grow”
❌ “Software companies that help eCommerce stores increase revenue through automated marketing”
💡 Sweet spot: 5–9 words
3. Match the language of your target market
AI reads real company descriptions.
Use:
“Software”
“Marketing agency”
“Consulting firm”
Avoid:
“SaaS”
“Solutions provider”
Internal jargon
Examples:
✅ “Software companies for eCommerce”
❌ “SaaS platforms for eCommerce”
4. Use prompt variations (always test)
Never rely on one prompt.
Run 3–5 variations per campaign.
Example:
Prompt variation | Companies found |
Marketing companies selling Facebook ads | 2,400 |
Marketing agencies selling Facebook ads | 2,800 |
Marketing agencies selling Meta ads | 3,100 |
Small wording changes = big volume differences.
5. Follow the proven prompt structure
Formula:
Company Type + Service/Product + Target Market (+ Outcome)
Examples:
“Accounting firms that help roofers”
“Cold email and lead generation agencies”
“Software companies selling to vets”
“Agencies helping contractors get more clients”
“Software helping eCommerce brands reduce cart abandonment”
📌 Think:
“Who they are + what they do + who they help”
6. Think like your buyer
Use outcome-based language found on websites.
Common phrases:
“Grow your business”
“Get more leads”
“Book more appointments”
“Close more deals”
Instead of:
❌ “Lead generation companies for home services”
Try:
✅ “Agencies helping home service businesses get more leads”
✅ “Companies that help roofers grow their business”
7. Test, don’t guess
Prompting is not intuition—it’s iteration.
Best practice:
Run multiple prompts
Compare company quality + volume
Save winning prompts for future campaigns
📌 Pro tip: Keep a “winning prompts” doc or spreadsheet.
What NOT to include in your prompt
Do not add filters inside your prompt.
Avoid:
❌ Locations (“in the US”)
❌ Funding stage (“Series A”)
❌ Job titles (“CMOs”, “CEOs”)
❌ Company size (“10+ employees”)
Why:
These already exist as separate filters
Prompting is for company identity, not attributes
Use:
Location filters
Job title filters
Funding filters
outside the prompt.
Prompting cheat sheet (save this)
Rule | Reminder |
🎯 Be specific | Niche > broad |
✂️ Keep it short | 6–9 words |
💬 Match language | “Software,” not “SaaS” |
🔁 Test variations | 3–5 prompts |
🧠 Think like buyers | Use website language |
🧩 Follow formula | Type + Service + Target + Outcome |
Expected outcome
You should now be able to:
Write high-performing AI prompts confidently
Generate larger, cleaner, more relevant company lists
Skip manual keyword and industry filtering
Build repeatable outbound workflows faster
Need help?
Use in-app support chat (bottom right)
Email [email protected]
Our team will help refine your prompts and improve results.
Prefer a video walkthrough?
Watch François (Head of GTM) break this down step by step inside the app.