You do not click through templates, campaigns, and automations to get work done. You talk, and the agent can make templates, build campaigns, set up automations, and work audiences. The dashboard is still there when you want it. Sends still wait on your rules.
2,000 most recent · 400/day · 15% holdout. Nothing sends until you approve.
Hey. You signed up for Acme back in March and never finished setting up. That's usually our fault, not yours.
So I did the step you stopped at. Your workspace is ready and there is nothing left to configure.
If it's useful, you can pick up where you left off today.
Same thread that can open a GitHub pull request. You do not have to click around the product to get the work done.
Ask for a welcome, a trial-ending note, a receipt. It drafts in the thread and publishes when you say so.
Audience, angle, holdout. It drafts the campaign in the same window and asks before anything sends.
Promote a campaign into an always-on automation, or ask it to build the flow from scratch.
Describe who should get the send. It builds the segment from product events.
Shipped features, stalled cohorts, campaigns ready to scale, overlaps with automations you already run. Each row gets a call: draft, narrow, hold, decline, or ask you.
| What the scan found | Ranla's call |
|---|---|
53,000 people signed up and never activated. None of them have ever been emailed. Wrote the campaign and put it behind a ramp, so it opens at 400 a day to the most engaged rather than going out to all 53,000 at once. | Drafted |
Three user-visible changes shipped since the last product update 12,400 people use the parts that changed, but only 4,000 sends are left under this week’s cap. Going to the best 4,000 now, the rest follow next week, rather than holding the whole thing. | Narrowed |
Variant B is ready to scale It earned 2.4× variant A over 900 sends. Rolling it out to the remaining 6,700 is waiting on your yes, because that is how you set this account. | Needs you |
Onboarding drop-off at day three An automation already covers the onboarding stage. If it is underperforming, the fix is to change that automation, not to stack a one-off campaign on top of it. | Declined |
Re-engagement push to everyone quiet for 60 days “Launch week recap” reached most of this audience four days ago. Holding a week so nobody hears from us twice. | Held |
Sends on the calendar, audits queued, performance reviews scheduled. You set how much it can do without asking. Anything past that limit waits for your yes.
Decisions from last week, proposals waiting, one intake question
Ramp day 3 of 5 · 400 sends today
Copy still describes onboarding from six months ago
Shipped checkout redesign · narrowed to 4,000 under weekly cap
Holdout results for finish-setup campaign
Scale variant B to the rest of the cohort?
DM it, @mention it in a channel, or use /ranla. Brief it, approve a send, see what it is doing. Open the app when you want the canvas. Slack does not replace the dashboard.
DM Ranla, @mention it in a channel, or run /ranla. Same agent that drafts in the product window.
Tell it what to do next. Approve a send, snooze a week, or reject, without leaving Slack.
The dashboard is still there when you want the full picture. Slack is how the team talks to it day to day.
When someone replies to a campaign, Ranla classifies what they want: pricing, how-to, unsubscribe, complaint, and drafts an answer in your voice.
Safe questions can go out on their own. Refunds, churn, and anything sensitive land with a draft ready for you to send.
Ranla connects a GitHub repo, scans the codebase, instruments the events the platform needs, and opens a pull request. You review and merge. Nothing ships until you do.
Connect GitHub and pick the repository. That is the only setup step on your side before the scan.
The agent reads how you sign people up, what you already track, and where the gaps are.
It wires the moments Ranla needs for audiences and automations, using the conventions already in the repo.
You get a PR. Merge it when it looks right. The agent does not push to your default branch.
Lifecycle email and product instrumentation in one platform. Marketing and engineering do not need two tools for that.
signed_upWiredactivatedWiredinvited_teammateRecommendtrial_endingRecommendOpens a PR when the repo is connected
Missing activation? A trial-ending moment with nothing behind it? The agent recommends events from the product. With the repo connected in the workspace, it instruments those events and opens a pull request. Campaigns and code changes from the same place.
Analytics connect (PostHog, Amplitude, Mixpanel) is still there if you would rather not change the repo.
Frequency caps, daily send limits, and audience-size approvals are enforced in the send path. The agent can tighten its own limits; only you can loosen them. Same rules in Slack and in the app.
Same agent that drafts and asks. Your other agents talk with it. Send rules still hold.
https://mcp.ranla.ai/mcpRanla includes a built-in agent that drafts campaigns and automations, then asks before anything sends. Add it to Cursor, Claude, Grok, or any MCP client. The agents talk to each other. Ranla still drafts and asks.
The in-product conversation where you brief Ranla: templates, campaigns, automations, audiences, and GitHub instrumentation. It drafts and asks. You approve sends in the app or Slack.
Yes. Connect GitHub. The agent scans the repo, recommends events, instruments them, and opens a pull request. You merge. It does not push to your default branch.
Yes. DM it, @mention it, or use /ranla. Brief it, approve a send, see progress. The canvas stays in the app when you want the full picture.
Connect a GitHub repo, or analytics. Then talk in the agent window: templates, campaigns, automations. It still asks before it sends.