What agents does Cleanlist offer?

Three: Research (investigates a contact, company, or both on the live web — 3 credits for a single scope, 5 for both), Qualification (scores a lead against your ICP and checks whether they're already in your CRM — 5 credits), and Custom (your own saved, reusable prompt — 3 credits at 500 resolved prompt tokens or fewer, 5 above).

What is the Organization Profile?

A structured go-to-market summary of your company that Cleanlist builds from your website and, optionally, your company LinkedIn page. It feeds the Research and Qualification agents so their output is grounded in what you sell and who you sell to. Building it is free, and one teammate completing it unlocks agents for the whole organization.

How do I save a reusable agent?

Write a Custom agent prompt, then save it as a named Custom Agent. It appears as its own card in the catalog and can be re-run on any list. Saved Custom Agents are org-wide and only admins can create, edit, or delete them.

What plan do I need to run agents?

Starter or above on self-serve, and every AppSumo tier. Agents are not available on the Free plan.

Why did my Custom agent return blank or 'unknown' for every row?

Almost always a prompt-variable mistake. Custom agents resolve display-label references like /Company Name, not snake_case like /company_name, and they do not resolve {{ }} tokens at all. An unmatched reference becomes an empty string silently, at full price. Test on 5 rows before scaling.

👤

Who this is for: self-serve plans (Free, Starter, Pro, Scale). If you bought a Cleanlist AppSumo lifetime deal, your features, limits, and pricing are different — see the AppSumo documentation instead. Support agent: confirm whether the user is self-serve or AppSumo before giving plan-specific advice.

TL;DR: Build your Organization Profile first (free, one time, unlocks the whole org). Then run Research, Qualification, or a Custom agent from the Agents tab of the Actions drawer. Agents need Starter or above, and are on every AppSumo tier. See Agents for the full reference.

Using Cleanlist agents

Agents work your leads for you: reading the live web, scoring against your ICP, and running prompts your team writes once. This guide walks through setup and the three run flows.

Agents are available on Starter and above on self-serve, and on every AppSumo tier.

Step 1: set up your Organization Profile

Research and Qualification are grounded in an Organization Profile — a structured summary of your company. Until someone completes it, the Agents page shows the setup card instead of the catalog.

Provide your URLs

On first use of the Agents page, enter your company website and, optionally, your company LinkedIn URL.

Let Cleanlist research your company

Cleanlist reads your homepage and runs deep web research to draft the profile: your products, positioning, ICP, personas, competitors, and geography. This runs synchronously and can take up to four minutes.

Review and approve

Review the draft. Edit any field that needs correcting, then approve it. Approval unlocks agents for everyone in your organization — onboarding only needs to be completed once per org.

Fill in the exclusion lists

Add job titles you never sell to under excluded personas, and any countries, states, or cities you do not sell into under the geographic exclude lists. These are the only fields that can disqualify a lead before any model runs, which makes them the cheapest filter you have.

Keep it current

Admins can edit the profile any time from the Profile tab, or re-run the research to regenerate it. Re-running is always free, capped at 6 rebuilds per hour per organization.

First-time setup is open to any org member so a teammate can bootstrap it. Re-running is admin only, because it rewrites context that flows into every teammate's agent runs.

Step 2: run an agent on a list

Open the list and the Agents tab

Open a lead list, click Actions, and switch to the Agents tab. You'll see the three built-in agents plus any Custom Agents your org has saved.

Pick your target rows

Choose selected rows, all unfilled rows, or re-run everything. Add an Only run if condition to skip rows that don't match. Turn on auto-update if you want rows added later to be processed automatically.

Choose scope (Research only)

Company (3 credits) — firmographics, positioning, signals, funding, hiring
Contact (3 credits) — role, tenure, prior companies, public content, talking points
Both (5 credits)

On the Companies sheet the scope is locked to Company.

Check the cost, then run

The panel shows the exact credit cost before you run. Results stream back row by row. Each row moves through Pending → Processing → Completed / Partial / Failed.

Runs are capped at 20 per minute per organization. A large list is queued and worked through in the background; you don't need to keep the tab open.

Step 3: read the results

Open any Research cell to see three things:

The written analysis
A Structured insights panel, where you can promote any single field to its own column
An Agent trace showing each search and page fetch the agent performed, redacted of personal data

Qualification writes back a fit_score (0-100), a verdict (qualified, disqualified, needs_review), 1 to 4 verdict_reasons, a crm_status, and any disqualification_flags.

Writing prompts that actually resolve

This is where credits get wasted. The variable syntax is not the same across the three agents.

Research and Qualification resolve {{namespace.field}} tokens. Type / in the prompt editor and pick from the list — the editor inserts the right token for you.

Custom agents resolve a completely different syntax: a forward slash plus the field's display label, capitalized, with spaces.

Identify which CRM /Company Name (/Company Domain) appears to use.
Reply with exactly one of: hubspot, salesforce, pipedrive, zoho, other, unknown.
Output only the value, no explanation.

/company_name is not valid. It resolves to an empty string, silently, and the run still costs full price. So do {{ }} tokens on a Custom agent — they are passed to the model as literal text. Always test a new prompt on 5 rows and read the output before scaling it.

The valid Custom agent labels are: First Name, Last Name, Full Name, Job Title, Email, Phone, LinkedIn URL, Company Name, Company Domain, Industry, Company Size, City, State, Country, Full Location, Created Date, plus any custom column on your list.

Saving a Custom Agent

Write a prompt, then save it as a named Custom Agent so your team reuses one vetted version. There's a Generate button that drafts a prompt from a plain-English description of what you want.

Creating, editing, and deleting Custom Agents is admin only. Deleting one removes the saved prompt but keeps every result it already produced.

You can also save a one-off prompt without spending a run, using Save settings only in the run panel.

Best practices

Get your Organization Profile right first. Research and Qualification quality depends on it.
Test on a subset. Run any agent on 5 to 10 leads before committing credits across a large list.
Filter before expensive runs. Narrow the list to rows worth scoring, then run the 5-credit Qualification or Research only on those.
Qualify, then research the winners. Qualification at 5 credits on everyone, then Research at 5 only on verdict = qualified, is far cheaper than researching a whole raw list.
Watch Custom prompt length. A resolved prompt over 500 tokens costs 5 credits instead of 3. The hard ceiling is 8,000 tokens.
Use exclusions. A lead caught by an exclusion rule skips the model entirely.

Agents — the full reference
Agent recipes — four working setups
Using Cleanlist Copilot — drive agents in plain language
List actions — validation and enrichment columns
Credit pricing

Managing Lead Lists Using Cleanlist Copilot