How it works
An automation belongs to a campaign. It watches a lead source, hands every new lead to the campaign — which runs the research and writes the sequence — and pushes the finished output to wherever your reps work. With Outreach as the source, the automation watches one or more tags in your Outreach workspace and pulls anything new it sees, on its own, forever.
The trigger is the tag. Apply a tag to a prospect in Outreach and that's what wakes the automation up. Twain polls the tag every minute, picks up prospects that weren't there last time, runs them through the campaign, and notifies the channel you've wired up when each one is ready.
Two practical points before the setup:
- You apply the tag, Twain handles the rest. Twain reads tags, it doesn't create them. Your existing Outreach tagging hygiene — territory, persona, intent signal, hand-raisers, whatever you use — becomes the queue.
- Tags are also the off-switch. Remove the tag from a prospect in Outreach and they stop being a candidate for that automation. They don't get retro-removed from a campaign already running, but they're out of the new-prospect queue.
Pick the tags that trigger the automation
There are two ways in, and both end up in the same place.
From the leads step. When you create a campaign, duplicate one, or add leads to an existing one, pick Outreach as the source and choose the tag. Below the tag list sits an Auto-add new leads switch: turn it on and the campaign keeps pulling from that tag after the one-off import is done. Finishing the campaign creates the automation, already running.
From the campaign. Open the campaign and find the Automation row in its sidebar, under Agent. Setup opens the builder on the campaign: an Import step for the source, an Export step for the destination, and a Notification step. Once an automation exists, the row shows whether it is Active, Paused, or has an error, and its gear opens the automation's diagram and settings in place.
Either way, the tag picker is the same. Search for the tag, click it. You can pick more than one — Twain treats them as an OR: any new prospect carrying any of the selected tags will flow through.
A couple of things to know about the picker:
- The counts are live from Outreach. They reflect prospects with that tag at the moment you opened the picker. They're a sanity check, not a frozen number.
- You can change the selection later. Twain re-uses the new tag set the next time it polls. Prospects already mid-flight finish on the rules they entered with.
Choose the research depth
Every lead the automation pulls in is generated by the campaign it belongs to — same agent, same persona, same sequence structure. What you decide at setup is how deep each generation goes. The research depth sits right next to the source, priced per lead for this campaign's step count:
- Base — 1 credit per lead.
- High — 2 credits per lead.
- Ultra — 8 credits per lead plus 2 per sequence step, and the deepest research and reasoning Twain runs.
The depth is a setting for leads that have not been generated yet, so you can change it later from the automation's settings; leads already mid-flight finish at the depth they entered with. On the automation's diagram it reads as one step — Generate in Base, High or Ultra — between the source and the export.
One thing worth calling out: the campaign is the automation's home, not just its destination. If the campaign was created from a different lead source, that source is what the one-off import used; the automation is the source of truth for everything that arrives after it is switched on. The campaign provides the agent, the persona, the sequence structure, and the credits.
Export onward
The Export step is where the finished output goes. It connects to the same destinations Twain supports anywhere else — Apollo, HeyReach, HubSpot, Instantly, La Growth Machine, Lemlist, Outreach, Salesforce. Pick the one your reps will work the lead in and map the sequence steps to the right fields on the other side.
A common pattern with Outreach-triggered automations is to export back into Outreach — different sequence template, but the same workspace your reps already live in. That way the tag-to-tag round trip stays inside one tool from the rep's perspective: they tag a prospect in Outreach, and a few minutes later that prospect shows up in Outreach again with a Twain-generated sequence attached, ready to launch.
You don't have to export to Outreach, of course. Plenty of teams import from Outreach and export to HubSpot (so marketing keeps a sequenced touchpoint in the CRM record), or export to Salesforce (so the activity rolls up against the opportunity), or to Lemlist / Instantly (if Outreach is the source of truth but a different tool handles delivery). Pick whatever matches how your reps actually work.
Putting it together
The end result is an automation that runs on its own. New prospect gets tagged in Outreach → Twain notices on the next poll → adds them to the campaign → runs research → writes the sequence at the depth you picked → exports to wherever you pointed the Export step → notifies you when it's done.
Once it's active, the only thing the rep does is tag prospects in Outreach. Twain handles the import, the research, the writing, and the delivery. The campaign's log is where you go to see what ran and when — every lead the automation added is signed by it — useful when you're tuning the campaign or debugging a tag that shouldn't have matched. In the campaign list, an automated campaign carries a mark that turns while a run is in progress, and the list can be filtered down to automated campaigns only.
For the import side of the same integration — campaign creation, one-off leads import, and how the tag picker behaves — see Import Outreach Prospects Into Twain.