Skip to main content

​
Outbound hub orientation — every tile and where it lands

The Outbound docs split coverage by feature — a guide for the create wizard, one for the audiences console, one for templates, one for AI Studio — because that is how the features shipped. Operators do not work that way. The question that brings someone to /outbound is one sentence: which tile do I start on? This page answers it, collecting every section on the Outbound hub in the dashboard into one walkthrough. Every section below is a tenant-owned control — you set it for your workspace, and nothing applies until you turn it on. Each tile maps to a route under /outbound in the dashboard.

​
Why the hub is a map

The Outbound hub stacks six sections top to bottom: a hero band with the create-send chooser, a 30-day performance strip, a quick-prompts card that routes into AI Studio, an in-flight and scheduled campaigns table, a deep-link grid to the sub-surfaces (Audiences, Templates, Goals, AI Studio, Route Preview, Approvals, OTP approvals, plus Flows and Ads), and a final attention row pairing recently sent campaigns with the ones that need a decision. One question brings an operator here — which tile do I start on? — and the hub answers it in this order: measure, launch, drill in.

​
Land on the hub — hero band and the create-send chooser

The hero band holds the hub title and one primary action: create send. Opening the chooser is the first decision the hub asks for — what kind of send you are about to run. The five options route straight into the right surface: Blast and Drip open the campaign create wizard; the wizard walks channel, audience, body, schedule, and preflight in sequence. Journey opens the journey builder. Direct send and Voice broadcast open their own surfaces. Docs: Campaign create wizard; Direct send; Voice broadcasts.

​
Read the 30-day strip — Sent, Delivered, Delivery, Cost

Directly under the hero, four cards answer how is my outbound traffic trending? over a rolling 30-day window:
  • Sent — outbound messages handed off the platform, excluding rows still queued or in flight.
  • Delivered — messages confirmed delivered in the same window.
  • Delivery — the delivered-to-sent ratio, the cross-channel engagement proxy until a replies-that-converted rollup ships.
  • Cost — total spend in your workspace billing currency.
Trust the trend when the sent volume behind it is non-trivial — a delivery rate computed over a handful of sends swings day to day. Cards render zero values when you have legitimately sent nothing, and each zero-state card links to the create wizard. When a card shows an em-dash with an inline retry banner, the metric read failed; retry, and treat the strip as advisory until it reloads.

​
Quick prompts — the AI suggestions card

The card below the strip offers three starting points, each with a one-line rationale, and each opening AI Studio with its context pre-loaded:
  • Resend to non-deliverers — carriers retry transient failures, and a fresh attempt with a tweaked body often wins.
  • Try the winning variant with a bigger audience — when an experiment declares a winner, scale the proven message past the test cohort.
  • Re-engage contacts with no opens in 30 days — a cooling-list segment plus a short reactivation prompt usually recovers a slice of the inactive base.
These are deterministic heuristics, honest about what they are — the card header reads Quick prompts. Each prompt routes into AI Studio (/outbound/ai) where the full generator runs. What feeds the generator decides the output quality: run it on the engagement and contact data your workspace already has, and power the inference with a credential you own — the workspace LLM provider credential (owner/admin-managed) or a default provider tier. Docs: Outbound AI Studio; BYO LLM inference credential.

​
In-flight and scheduled table — cancel a queued send, open a live campaign

The compact table merges two campaign states into one list: running (currently dispatching) and scheduled (queued to send), sorted by scheduled time so the next-to-fire row sits at the top. Each row shows the campaign name, channel, audience size, scheduled time, and live send progress (sent of total) for running rows. Click a row to open the campaign detail — the working surface for a live campaign (progress, per-recipient outcomes, per-variant stats). For a scheduled row, the detail page is where you cancel the queued send before it fires. Rows awaiting sign-off show as pending approval and route to the approvals queue. The View all link opens the full campaigns list with every lifecycle state. Docs: Outbound goals, approvals, and direct send; Message scheduling. The secondary grid is the permanent map into every Outbound sub-surface. Each tile deep-links to the console that owns the job:

​
The attention row — recent results and failures without a deep dive

The bottom row splits into two equal lists. Recently sent shows the last five completed sends, so yesterday’s launch is one click from its results. Needs attention lists recent campaigns that ended failed or partially failed, with the failed-recipient count per row. Read it as a signal, not an investigation station: a row here means some recipients did not get the message — suppressed numbers, frequency caps, quiet-hours deferrals, or provider rejections. Open the campaign detail only when you decide the counts warrant a resend; the quick-prompts card above suggests exactly that path. An empty attention list is the healthy state — nothing needs a decision, and the list says so instead of nagging.

​
First login — the empty-state checklist

On an organization with no outbound history the hub is still readable, and reads as a checklist:
  1. The 30-day strip shows zero-value cards, each linking to the create wizard — the hub steers you to a first send instead of showing a blank chart.
  2. The quick-prompts card stays useful on day one because its prompts are deterministic starting points; the full generator behind them deepens once real engagement data exists.
  3. The in-flight table renders an empty state with a Create a campaign action.
  4. The deep-link grid renders in full — Audiences, Templates, Goals, AI Studio, and the rest are reachable before any send exists, so you can build the audience and the template first.
  5. The attention row shows its positive empty states — nothing sent yet, nothing needing attention.
The standing order that works: stand up an audience and a template from the grid, launch a first blast from the chooser, then let the strip and the attention row start carrying real numbers.

​
Role gating and tenant isolation

The hub renders for owner, admin, and operator seats. The two approval queues — campaign approvals and OTP approvals — are owner/admin-gated at the destination page, even when the tile is visible. Every surface here is tenant-owned: campaigns, audiences, templates, and approval queues live in your workspace only, nothing crosses tenant boundaries, and nothing applies until you create it. Outbound runs under the send controls you configure — suppression lists, frequency caps, and quiet hours all gate before a dispatch leaves the platform.

​
See also