Decide

Build the journey on a canvas. Simulate it before it runs.

Triggers, waits, branches, goals and channel actions on a visual canvas, with per-node stats, versioning and a 30-day replay.

journeys / abandoned-checkout · v3 · draftSimulateSubmit for approval
yesnoTRIGGERcheckout_startedno order_completed in 45 min4,318 enteredCONTROLControl group10% hold back432 heldACTIONWhatsAppcart · live price3,674 deliveredWAITWait 4 hexit on order_completed1,102 exited · goalBRANCHOpened?channel response2,572 evaluatedACTIONPushtop item · deep link1,140 sentACTIONEmail · 09:00 localrecommendations1,432 sentGOALorder_completedconversion window 24 hattributed vs control

Overview

A journey is trigger, conditions, ordered actions and a run history that stays explainable after the journey is edited. Edverix adds the two things every marketer wants before pressing activate: a simulation of the last thirty days of real events against the draft, and a version that lives in git if your team prefers code to canvas.

Only on Edverix

  • LLM stepOnly on Edverix

    Pass customer and event context to a model mid-flow and use its output to write, classify or choose a branch.

  • Simulate against thirty days of historyOnly on Edverix

    See exactly who would have entered, what they would have received and how often, before activating.

  • Run history with explainabilityOnly on Edverix

    Every run snapshots the evaluated state, so it stays explainable after the journey is edited.

Capabilities

Everything in Journeys

16 capabilities in 4 groups, organised the way the product is.

Triggers

Event, schedule, segment, attribute, date, geofence, API

Start a journey from almost anything, including a call to the transactional endpoint.

Business-event relays

A price drop, a restock or a new episode fans out only to customers whose behaviour shows affinity. One setup replaces hundreds of campaigns.

Contextual journeys

Independent trips per order, policy, loan or enrolment, so a second purchase never collides with the first.

Flow control

Wait for a period, an event, a date or a window

"Wait up to 48 hours for purchase, else send the reminder" is two nodes.

Branch on anything

User attribute, event property, random split, best channel, or what the last message did: sent, delivered, opened, clicked, failed, unreachable.

Exit triggers and goals

Drop a person out the moment they convert, so you never chase someone who already bought.

Write back to the profile

Tag participants, set a status or increment a counter mid-journey.

LLM stepOnly on Edverix

Pass customer and event context to a model mid-flow and use its output to write, classify or choose a branch.

Safety and proof

Simulate against thirty days of historyOnly on Edverix

See exactly who would have entered, what they would have received and how often, before activating.

Per-node statistics

Entries, flow-through and drop-off at every node, aware of which version each person went through.

Run history with explainabilityOnly on Edverix

Every run snapshots the evaluated state, so it stays explainable after the journey is edited.

Maker-checker approval

The person who builds a journey is not the person who approves it.

Journeys as codeOnly on Edverix

Define a journey in your repository, review it in a pull request, validate in CI, promote from staging and roll back like any deploy.

Campaigns

One-time and recurring campaigns

A single send to a segment, now or scheduled, with A/B variants and a control group.

Transactional sends

An API endpoint that bypasses frequency caps and quiet hours, because a receipt or an OTP is never marketing.

Deliver in local time

One campaign, sent at nine in the morning wherever each person actually is.

FAQ

Questions about Journeys

What happens to people mid-journey when I edit it?

You choose: finish on the version they entered, or migrate them to the new version at their current node. Either way the run history records which version each person saw.

Can a journey call our own systems?

Yes. A webhook action posts a signed payload to any endpoint, and a wait-for-event node can wait on the event your system sends back.

How is the simulation different from an audience estimate?

An estimate is a count. The simulation replays real events through the draft and returns the exact people, messages and timing, including where a cap or quiet hours would have stopped a send.

Get started

See Journeys on your own data.

Send your first event in an hour, simulate your first journey against real history, and keep every byte of it in a database you can query.

  • No credit card for the Developer plan
  • Bring your own providers
  • Export everything, any time