To build a lead list with AI, write one sentence about your buyer, check what the tool understood and how many people match, then pay only for the leads, emails and mobiles you keep. This walkthrough takes one sentence to a 100-person call list in Scalelist, with the credit cost of every step and the numbers from one of our own runs.
In our July 2026 test, the search, the list and the emails took a few minutes in one browser tab. Mobiles and the export are two more actions, so plan on about 10 minutes for your first list. Most of that time is you reading the results.
What does it mean to build a lead list with AI?
Building a lead list with AI means describing the people you want in plain English and letting a tool turn that sentence into named people with contact details. You check how it read your sentence before anything is charged, instead of translating your buyer into menus yourself. It is the first step of finding B2B leads with AI: the list everything else runs on.
A chatbot on its own cannot do this. ChatGPT has no verified contact database; it reaches contact data only through connected apps. An AI lead finder resolves your sentence against real people and verifies every email before you see it. For the chatbot side, read what ChatGPT can and cannot do for lead generation; for the menu side, our comparison of natural-language search and filter-based databases.
The routine at a glance
The routine has seven steps, and the first three cost nothing.
We will follow one example. Kate Morales is the founder of Rampwise, a fictional Austin, Texas startup that sells onboarding software for new sales reps. She wants a call list of VPs of Sales at software companies in the US.
How to build a lead list with AI in 7 steps
Here is each step in Scalelist’s AI lead finder, with what it costs and what to check before you move on.
Step 1: Write one sentence about your buyer
Write who you want the way you would brief a new rep: a title people actually hold, a company type in plain words, a location and a size range. Kate writes:
This step costs 0 credits. If you are unsure how to phrase your buyer, our guide on how to describe your ICP in one sentence covers the method, and our library of 60 prompts by role and industry gives you starting points.
Step 2: Check what Scalelist understood
Scalelist turns the sentence into chips: title, seniority, industry, location, company size and time in role, as far as your sentence mentions them. Kate’s sentence reads as four chips.
If a chip is not what you meant, for example a different industry, rephrase the sentence and run it again. This step costs 0 credits.
Step 3: Size the list before you pull anyone
The count tells you how many people match, and seeing it is free. On 27 September 2026, Kate’s sentence matched 1,774 people. The same title across all US companies matched over 145,000, and adding “who changed jobs in the last 3 months” to Kate’s sentence cut it to 89.
Pick the version that fits the week you have. Phone-centric SDR teams average 56 dials a day, according to The Bridge Group’s 2025 SDR report, so 100 people is about two days of first calls for one rep. The 89 new leaders make a sharp second list; see how to find new executives in their first 90 days.
Step 4: Add your first 100 leads
Add 100 people from the 1,774 to a list. Each lead added costs 1 credit, so this step costs 100 credits. The other 1,674 stay in the market for your next batch.
Before you spend anything else, read the first 25 rows. Remove companies you would never sell to. Your judgment on fit is still the best check.
Step 5: Reveal emails for the batch
Reveal emails for the whole list at 1 credit per email found. Every email is verified before it is returned and labelled Valid or Risky; Not found costs nothing.
Risky means the company’s mail server accepts every address, so the mailbox cannot be confirmed. About 20 to 30% of Risky emails are deliverable, so send them from a secondary domain. Our guide to catch-all emails covers how to handle them.
Step 6: Reveal mobiles for the people you will call
Reveal mobile numbers only for the people you plan to call this week, at 20 credits per mobile found. A mobile that is not found costs 0.
In our North America benchmark, Scalelist found a phone number for 87% of the list. At that rate, 100 people give Kate about 87 mobiles for 1,740 credits. A mobile rings the person, not a switchboard; here is why direct mobiles connect more often.
Step 7: Export the call list or push it to your tools
Export the list to CSV or Excel for your dialer, push selected leads to HubSpot with field mapping, or send them to lemlist, Instantly, Smartlead, Woodpecker or Saleshandy. This step costs 0 credits. For dialers, see how to send Scalelist mobiles to your dialer.
Name the list with the sentence and the date. Next month, you can run the same sentence again and pull the next batch.
What a 100-person call list costs
A 100-person call list with emails and mobiles costs at most 2,200 credits, and about 1,932 at our North America benchmark hit rates.
| Step | Credit rule | If everything is found | At benchmark rates |
|---|---|---|---|
| Count the market | Free | 0 | 0 |
| Add 100 leads | 1 per lead | 100 | 100 |
| Reveal emails | 1 per email found | 100 | 92 |
| Reveal mobiles | 20 per mobile found | 2,000 | 1,740 |
| Total | 2,200 | 1,932 |
The 2,000-credit plan costs 89 USD a month, so one mobile costs about 0.89 USD on it. If your hit rate is higher, unused credits from earlier months roll over (up to 2x your monthly credits), or step up to 5,000 credits for 199 USD. Every plan is on the Scalelist pricing page.
Size your market for free, then pay only for the emails and mobiles Scalelist finds.
Describe your buyer in one sentenceWhat a real run returned: our July 2026 test
We ran one sentence through Scalelist on a live account in July 2026: “CTOs of SaaS companies with fewer than 500 employees, only based in the Tampa Bay, Florida area.” Tampa Bay is a mid-size metro, where thin coverage would show.
- 148 people matched. Seeing the count cost nothing.
- 148 leads added: 148 credits.
- 117 emails found (79%): 102 Valid and 15 Risky, 117 credits. Risky emails were labelled, not hidden.
- 31 people with no email found: 0 credits.
- Total: 265 credits. We did not reveal mobiles in that run.
Your rates will differ by seniority, company size and region. Test a small batch in your own market before you plan a quarter on it. Our provider benchmark by region shows how coverage changes outside North America.
Tips for a better AI lead list
Good lead list building with AI comes down to narrow sentences, small batches and a human check before you spend on mobiles.
Keep one audience per sentence
“VPs of Sales and CFOs at software and manufacturing companies” mixes four lists. Write one sentence per audience, so each list gets its own talk track.
Pull in batches you can work this week
Contact details go stale while a list waits. Pull what you can call in a week, work it, then pull the next batch from the same sentence.
Recount before every new batch
Counts move as people change jobs. Recounting is free, so check the number before you pull the next 100.
Spend on mobiles last
Emails cost 1 credit and mobiles cost 20. Reveal mobiles after you have removed wrong-fit companies, never before.
Troubleshooting
Most problems with an AI lead list show up in the chips or the count, before you spend a credit.
The count is too big
Add a size range, a state or a city, or a narrower title. “Sales leaders” is broad; “VPs of Sales” is not.
The count is small or zero
Check the chips first: a 0 often means a word was read differently than you meant. Then widen one thing at a time: add a common title variant (“Head of Sales”), move from a city to a state, or widen the size range.
A chip is not what you meant
Rephrase with plainer words. Write the industry the way a trade publication would (“software”, “staffing and recruiting”) and run it again. It costs nothing.
Many emails come back Risky
Risky emails come from companies whose mail server accepts every address. Send them from a secondary domain, or call first and email after the conversation.
Some people have no mobile
You are not charged for them. Email them, or call the company’s main line and ask for the person by name.
Turn one sentence into a call list with verified emails and direct mobiles. Misses cost 0 credits.
Get 20 credits to test itFAQ
Can ChatGPT build a lead list?
Not on its own. ChatGPT can help you describe your buyer, but it has no verified contact database and reaches contact data only through connected apps. Connect a data tool, such as the Scalelist MCP server, or use an AI lead finder that resolves your sentence against real people and verifies emails before you see them.
How long does it take to build a lead list with AI?
Plan on about 10 minutes for a first list of 100 people. In our July 2026 test, the search, the list and the emails for 148 people took a few minutes in one browser tab. Revealing mobiles and exporting are two more actions. Most of the time goes into reading the rows and removing wrong-fit companies.
How much does a 100-person call list cost in Scalelist?
At most 2,200 credits: 100 for the leads, up to 100 for emails and up to 2,000 for mobiles. You pay only for what is found, so at our North America benchmark rates the same list costs about 1,932 credits. That fits the 2,000-credit plan at 89 USD a month.
What should I do if the count is zero?
Check what Scalelist understood first. A zero usually means a word in your sentence was read differently than you meant. Rephrase the industry in plain words, add a common title variant, or widen the location from a city to a state. Each recount is free, so change one thing at a time until the count makes sense.
Can I build the list inside Claude or ChatGPT?
Yes, through the Scalelist MCP server. Add the connector in Claude or ChatGPT, then ask for a count first, a batch of leads second and contacts last. Credits work the same as in the app. Our guide to prospecting inside Claude shows the prompts and the setup.
Is it legal to cold call the mobile numbers on my list?
In the US, the Telephone Consumer Protection Act (TCPA) rules restrict calls to mobile phones made with an autodialer or a prerecorded voice unless the person consented (47 CFR 64.1200). Dial by hand and keep your own do-not-call list. Rules differ by country; our guide to B2B cold calling laws covers the US, UK, EU and Australia. This is not legal advice.
Sources
- The Bridge Group: Sales Development Models, Metrics and Compensation, 2025 Research Report
- eCFR: 47 CFR 64.1200, delivery restrictions (TCPA rules)
- Scalelist people count (POST /api/v1/people/count), queried 27 September 2026, 0 credits.
- Scalelist test run on a live account, July 2026.
