Get pricingBook a demo

Planning a campaign

Every campaign can have a plan: a proposed list of the emails, pages, content, social posts, and automations needed to hit its goal. You don't have to build this list by hand — describe what you're trying to achieve, and the assistant can generate a full plan for you to review, adjust, and approve piece by piece.

Steps

  1. Open the campaign and go to its Settings tab.
  2. In the Goal section, describe what you're trying to achieve in plain language — for example, "drive 200 demo bookings from our email list before the quarter ends." When you click away from the field, the assistant may suggest a specific measurable target (a metric, a value, and a time window) based on your wording. Click Adjust to apply the suggestion, or leave your own wording if you'd rather set the target yourself later.
  3. Fill in the Brief section — what you're promoting, who it's for, and any offer involved — and choose a campaign type. You can browse example briefs for inspiration if you're not sure where to start.
  4. Once a campaign type is set, click Generate campaign near the top of the Settings tab. This opens a chat with the assistant, already primed to propose a plan grounded in the goal, brief, and type you just filled in.
  5. The assistant proposes a full plan, grouped by kind — emails, pages and forms, content, social posts, audience segments, and automations — so you can review it as a whole rather than one item at a time.
  6. Back on the Settings tab, scroll down to the plan list that appears below your settings. Go through each item and check or uncheck it to approve or exclude it from generation.
  7. Click Generate approved items to have the assistant build out everything you approved at once, or generate a single item by opening it and starting it individually.
  8. Click any plan item to open its details — what it is, its current status, and a link to the chat session that's generating or generated it. An item marked Needs you or Failed also shows the reason it stopped there, so you can see what to fix without opening the session.
  9. Needs you means generation stopped on a question it can't answer for you — most often a real-world fact that isn't written down anywhere, like the date and time of the event a journey is built around. The reason shown is the question. Answer it by editing the item's instructions in the same panel to include the fact (for example, replacing "one week before the event" with the actual date and time), then click Generate on that item to pick up where it left off. If you're answering in the chat session instead, the assistant writes the answer into the instructions for you and restarts the item — you don't have to do both.
  10. If you think of something the plan is missing, use Add item from your experience to describe it to the assistant in a new chat rather than adding a blank item yourself.

Example prompts

Generate campaign primes the chat for you, so you rarely have to type this yourself. If you'd rather ask directly, or want to add to a plan later:

  • "Generate a campaign plan for this campaign. Propose the full list of items — emails, pages and forms, content, social posts, and automations — grounded in the campaign's goal, brief, and type. Present them as a summary; I'll review, approve, and generate them in Settings."
  • "The plan is missing a reminder email the day before the event. Propose it as an additional item with instructions detailed enough to generate from."

That last phrase matters more than it looks. Each plan item carries a set of instructions, and those instructions are the only context the session that eventually generates that item receives — it doesn't see this conversation. The more specific they are (audience, goal, subject angle, CTA, tone, constraints), the less that later session has to stop and ask you. If you're adding an item by hand, write its instructions the same way you'd write a good prompt.

Tips

  • Automations (journeys and routines) are always created as drafts — review and turn them on from the automations page once they're generated. A routine needs instructions before it can be activated; Activate stays disabled until you've told it what to do. Once a routine is on, Run now kicks off an immediate run and shows Running… until that run finishes.
  • Re-running Generate campaign clears out suggested items you haven't approved yet before proposing a fresh plan, so you won't end up with duplicates of things you never approved. Items you've already approved are kept, even if they haven't been generated yet.
  • You don't need a plan to use a campaign — you can skip straight to chatting with the assistant about individual pieces of content if you'd rather work that way.

Generating a journey before its emails

A journey step that sends an email points at an email template. If the journey is generated before its templates exist, those steps are created empty — the journey looks complete on the canvas, but every send has nothing to send.

So when you press Generate and the batch contains a journey whose templates haven't been generated yet, you'll see a summary of what's missing and two choices:

  • Generate those first — close the dialog, generate the templates, then come back to the journey. Everything is wired automatically.
  • Generate anyway — build the journey now with those steps left blank.

Neither is wrong. Generating the journey first is a common way to work: you get the shape of the sequence, then write the emails to fit it. If you take that route, you don't have to wire it up by hand afterwards — once the templates are generated, the journey picks them up without being regenerated, and the sequence you designed is preserved.

Whichever order you work in, each email template, form or landing page you generate is recorded automatically against the journeys in the campaign that haven't been generated yet. You'll see it appear at the bottom of a journey item's instructions under Generated campaign artifacts — that's how the journey knows which template belongs in which step, so you don't have to pick them one by one on the canvas.

You'll see the same summary whether you're generating one item or the whole approved list; it counts across everything in the batch.

When an item needs your attention

Some items finish as needs input rather than a draft. That means the assistant built something it couldn't complete correctly, and the reason is shown on the item. Two common cases:

  • A journey with a step that can't work as described — for example a condition that could never be true, or a trigger for something the app doesn't track. The journey is still there and you can open and fix it.
  • A segment narrower than you asked for — if part of your description can't be expressed with the contact fields available, and you weren't told about it beforehand, the segment isn't saved and the item tells you which part was the problem. This is deliberate: a segment that quietly targets a smaller audience than you intended is worse than one that stops and asks.

An item in needs input can be generated again once you've adjusted it.

Audience segments: what you're told up front

When the assistant proposes an audience segment, it checks the description against your actual contact data straight away, before the item is ever created. Open the item and you'll see the result at the bottom of its instructions, under Grounded segment definition: the exact membership rule that will be built, and how many contacts match right now.

If part of your description couldn't be expressed, it's listed there too, under Not enforced by this segment — so you can see it while you're reviewing the plan, not after clicking Generate. The commonest example is timing measured from another action ("hasn't converted within 7 days of signing up"): a segment can filter on how recently someone opened an email or visited a page, but it has no record of when a tag was applied, so a window measured from that moment can't be enforced. That kind of rule belongs in the automation instead, and the assistant moves it there.

A segment with a Not enforced note still generates normally — the note is the disclosure, and the item builds to a draft you can review. If you'd rather it worked differently, edit the item's instructions and generate it again.

Zero matching contacts is normal and not an error: most campaign segments are defined on a tag that a form or automation applies later, so they start empty and fill up as the campaign runs.

FAQ

Q: I generated a journey before its emails. Do I have to start over? A: No. Generate the email templates, and the journey picks them up — the steps, waits and branches you already have are untouched.

Q: What happens to items I don't approve? A: They stay in the plan marked as excluded. You can change your mind and include them later — nothing is deleted.

Q: Can I add my own item to the plan instead of asking the assistant? A: There's no blank form for adding an item directly. Use Add item from your experience to describe what you want in a chat, and the assistant adds it to the plan for you.

Q: Do I need to set a goal before generating a plan? A: You don't strictly need one, but the goal and brief are what the assistant grounds the plan in — filling them in first produces a much more useful proposal.