The problem this solves
Onboarding email is where the gap between what a product team knows and what a new user hears is widest. The docs contain the real activation path - the setup step that unlocks everything, the feature that predicts retention, the mistake every new account makes in week one. The welcome sequence, meanwhile, says hello, lists five features with equal enthusiasm, and links to the docs it should have been distilled from. New users churn in the gap, having never been walked to the moment where the product proves itself.
The sequence stays generic because writing a good one requires an unusual combination: enough time inside the documentation to extract the true path to value, enough judgment to cut everything else, and enough craft to give each email exactly one job. Product knows the path but does not write lifecycle email. Marketing writes email but learns the product from the same marketing pages customers see. The synthesis falls between chairs, and the placeholder welcome series quietly becomes permanent.
Reading a documentation corpus, extracting the activation sequence, and drafting disciplined copy against it is synthesis work an agent does well - and the result should land as a built automation rather than a document, because a drip that exists only as a draft onboards nobody. The docs already paid for the knowledge; the mission turns it into the sequence new users actually receive.
How the mission runs
- The docs are read like a new user's map. Web Research works through your documentation - guides, setup instructions, FAQs - reconstructing the path from signup to first real value. It notes the dependencies: what must happen first, which steps users evidently stumble on given how much documentation defends them, and where the payoff moment actually lives.
- The activation path is proposed, five steps long. Before drafting, the agent shows you its read: the five moments that matter in order, and what it cut from the sequence. You correct the map where lived reality diverges from the written docs. Five minutes of correction here shapes every email more than any copy edit later.
- Five emails are drafted, one job each. Each email owns a single step of the confirmed path: do this one thing, here is why it matters, here is exactly how. Features that would distract are omitted entirely rather than mentioned briefly. Subject lines promise the step's payoff, and each email's success is one action taken.
- You review the sequence as a whole. The drip comes to you as a complete arc so you can judge pacing and progression, with each email's intent stated alongside its copy. Edits are applied across the set consistently. Nothing is created in your marketing platform until the whole sequence has your sign-off.
- The automation is built in ActiveCampaign, staged. On approval, the agent creates the automation in ActiveCampaign: the five emails, the signup trigger, and the agreed intervals. It reports what was built, linked for inspection, and leaves the automation staged so activation happens by your hand after a final platform-side check.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A five-email onboarding automation in ActiveCampaign built from your actual documentation: each email advancing one confirmed step of the activation path, in a disciplined arc you approved as a whole, wired to trigger on signup with agreed intervals and staged for your activation. Plus the extracted activation map itself, which tends to outlive the emails as a shared answer to what new users should do first.
Make it yours
- Branch by signup intent: users arriving from a specific use-case page get a variant sequence whose examples match that use case, drafted from the same activation map.
- Compress for trials: a 14-day trial gets the same five jobs on a tighter schedule, with the final email honestly addressing the convert-or-lapse decision.
- Add a stalled branch: users who have not completed the first step by email three get a variant that removes the friction it found documented most heavily.
Frequently asked questions
How does the agent know the real activation path rather than guessing?
It reconstructs the path from your documentation's own structure and defenses, then shows you the map for correction before drafting. That checkpoint is the accuracy mechanism: the docs propose, you confirm from lived data, and the emails are written against the corrected path rather than either alone.
What does the build touch inside ActiveCampaign?
It creates the new automation and its five emails, and nothing else changes: existing automations, lists, and templates stay untouched. The completion report names every created object, and the automation arrives staged rather than active, so no new signup receives anything until you turn it on.
Our docs are messy and partly outdated - will that poison the drip?
The map-review step is the filter: the agent flags places where the docs contradict themselves or look stale, and your corrections overrule the text. Outdated instructions get caught when you review the five-step path, which is far cheaper than catching them after a new user follows a wrong email.