Note: This video shows an earlier version of the lead view. Warnings now appear in the lead's details. Follow the written steps below.
What is Twain's Warning System?
It's a feature that uses AI to review your lead list and flag leads that might not be a good fit for your campaign, based on specific rules you define. It catches issues before you launch, so you're not spending credits or sending messages to the wrong people. On the website you may also see it called the qualification agent: same feature.
What you see depends on your plan
Warnings are generated for every lead on every plan. What is shown differs:
Starter: you see that a lead was flagged and the issue title (for example, "changed jobs"), but not the full findings behind it.
Pro and Enterprise: you see the issue titles and the full findings, including the evidence Twain found.
The same applies in the API and MCP: on Starter, responses indicate that a lead has warnings and their titles, and the detailed descriptions become available once you are on Pro. If you upgrade, the full findings for already-generated leads are shown immediately; in the API, request the lead again to receive them.
What kind of issues does it detect?
Outdated emails, people who changed jobs, non-decision-makers, inactive or career-break profiles, and leads who don't match your campaign's target criteria.
Where do I see the warnings?
In the campaign view, open any lead. Warnings are shown in the lead's details. Use the … menu next to a warning to Remove warning or Delete lead.
Can it check campaign-specific criteria?
Yes. If your campaign says "only reach out if they've mentioned Clay," Twain will flag leads that don't meet that condition. Warnings work against whatever rules you define, not just the defaults.
How do I set up warning rules?
Two ways:
Via the chat panel: Tell Twain what to flag. For example: "Flag leads who aren't in a decision-making role" or "Warn me if the company has fewer than 50 employees." Twain updates your Agent settings accordingly.
Manually: Open your agent, go to the General tab and edit the Warnings section (Qualification rules). Twain already flags common issues; add your own disqualifiers, risks or red flags there.
Does it use public data?
Yes. Twain can surface evidence from LinkedIn posts, webinars, or job ads to check whether someone or their company meets a condition, for example, whether they've used a specific tool.
How do I act on flagged leads?
Click the red warning icon next to your lead count to filter the list to warning leads only. From there you have three options:
Remove warning (per lead): keep the lead but dismiss the flag if you've reviewed it and it's fine.
Delete lead (per lead): remove the lead from the campaign entirely.
Delete X leads (bulk): a button at the bottom of the filtered view deletes all warning leads at once.
Warnings are grouped by category. You can also Remove warnings or Delete leads for a whole category at once.
Can I automatically skip flagged leads on export?
Yes. When you export a campaign to your CRM or sequencer, turn on Skip leads with warnings. The same option is available in an automation's Export step, so flagged leads never reach your sending tool. Note that research exports are available on Pro and Enterprise. See How Twain Automations Work.
Is this available outside the app?
Yes. Warnings also show up via the API when using Twain with Clay or custom setups, with the plan-based detail levels described above.