New: API Reference docs are live — integrate Cleanlist enrichment into your apps. View API docs →
Guides
Using AI Agents

TL;DR: Cleanlist has three AI Agents — Research, Qualification, and Custom — that run on any lead list. Onboarding auto-builds an Organization Profile from your website and LinkedIn, which grounds the Research and Qualification agents. CLAI, the in-app chat assistant, lets you drive the same work in plain language and always shows a confirmation card before spending credits.

Using AI Agents

AI Agents are Cleanlist's analytical layer. Where Smart Columns handle cheap, deterministic data hygiene, agents do open-ended research and qualification grounded in your company's go-to-market context. This guide covers the three agents, the Organization Profile that powers them, and CLAI.

The three agents

AgentWhat it doesScopeCost / lead
ResearchDeep-researches a lead's company and/or contactCompany, Contact, or Both3 (company), 3 (contact), 5 (both)
QualificationScores the lead against your ICP and checks whether they're already in your CRM5
CustomRuns your own prompt; output is markdown. Save reusable prompts as Custom Agents~3 (≤500 tokens) / ~5 (>500 tokens)

Research and Qualification also run as the research and qualification Smart Columns — adding the column on a lead list is the same as running the agent on those leads.

Set up your Organization Profile

Research and Qualification are grounded in an Organization Profile — a structured go-to-market summary of your company that Cleanlist builds during onboarding. It feeds your company's positioning, products, and ICP into every Research and Qualification run, so results are relevant to what you actually sell.

Onboarding is admin-only and free — it never costs credits, even when you re-run it.

Provide your URLs

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

Let Cleanlist research your company

Cleanlist scrapes your homepage and runs deep web research to draft a GTM profile — your products, positioning, and ideal customer profile. This runs synchronously and can take a couple of minutes for a thorough pass.

Review and approve

Review the drafted profile. 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.

Keep it current

Admins can edit the profile any time from the Profile tab, or re-run the research to regenerate it as your company evolves. Re-running is always free.

Run the Research agent

The Research agent gathers context about a lead's company, contact, or both.

Open a lead list and choose Research

Open any lead list and add the Research agent (it runs as the research Smart Column).

Pick a scope

Choose how deep to go:

  • Company (3 credits) — firmographics, positioning, recent signals about the account
  • Contact (3 credits) — background on the individual lead
  • Both (5 credits) — company and contact together

Select leads and run

Select the leads to process (test on a small subset first), confirm the credit estimate, and run. Results stream in as each lead completes.

Read the results

Each result is delivered as both a markdown summary you can read inline and structured sub-fields you can promote to their own columns.

Run the Qualification agent

The Qualification agent scores each lead against your ICP and checks whether the contact already exists in your connected CRM.

Add the Qualification agent

Open a lead list and add the Qualification agent (it runs as the qualification Smart Column). Cost is 5 credits per lead.

Run on your target leads

Select leads and confirm. Qualification uses your Organization Profile's ICP definition, so make sure your profile is accurate first.

Read the verdict

Each lead gets a fit score, a verdict, and reasons, plus a CRM presence check — a signal for whether the contact is already in your CRM. Filter and sort by the fit score to prioritize outreach.

Prefer Qualification over the cheaper icp_fit_analysis Smart Column when you want explained reasons or CRM context. Use icp_fit_analysis for a fast, low-cost score with no CRM lookup.

Run and save a Custom agent

Custom agents run a prompt you write. The output is markdown, and you can save a prompt as a reusable Custom Agent.

Write your prompt

Add a Custom agent and write a prompt describing what you want extracted or analyzed for each lead. You can reference lead, contact, company, and Organization Profile variables in the prompt.

Test on a sample

Run on a handful of leads first to confirm the output is what you expect. Cost is roughly 3 credits per lead for short prompts (≤500 tokens) and 5 for longer ones.

Save it as a Custom Agent

If the prompt is one you'll reuse, save it as a named Custom Agent (admin-only). It appears as its own card in the agent catalog and can be run on any list later. Each org's Custom Agents are private to that org.

CLAI — the in-app assistant

CLAI is Cleanlist's in-app chat assistant. Instead of clicking through the UI, you can ask CLAI in plain language to:

  • Search for people or companies
  • Import and enrich leads
  • Run Research, Qualification, or a saved Custom agent on a list

CLAI uses slash-command skills and chains steps naturally (search → import → enrich). Any action that spends credits shows a confirmation card before it runs, so you always approve the cost first — CLAI never claims an action is done before you confirm.

Best practices

  • Get your Organization Profile right first. Research and Qualification quality depends on it. Review the drafted profile carefully and fix anything wrong before running agents at scale.
  • Test on a subset. Run any agent on 5-10 leads before committing credits across a large list.
  • Filter before expensive runs. Score cheaply (e.g. icp_fit_analysis), filter to high-fit leads, then run the 5-credit Qualification or Research agent only on those.
  • Reuse with Custom Agents. Save prompts you run repeatedly so your team runs a consistent, vetted prompt instead of rewriting it each time.

Learn More