See exactly what happened to one customer in your Action Flow

When a customer tells you they didn't get the card they were expecting, where do you go to find out why?

For most teams, the answer is: a Slack to engineering, a query against a data warehouse, or a guess based on aggregate stats that can't tell you anything about one specific person. Something happened to that customer, on that flow, at that moment, and your tooling shows you a chart of conversion rates instead.

A new way to investigate

The new Runs tab in the Action Flow editor gives you a single place to inspect any individual run of an Action Flow, from trigger to completion. You can see which path the customer took, which conditions evaluated true, which cards were sent or withheld, and which steps failed or are still waiting.

It is the answer to "what actually happened to this person?"

How it works, in plain language

Open any Action Flow in the Workbench and click the Runs tab. You see a list of recent runs for that flow. Click one, and the Action Flow graph opens with the run overlaid on top, every step colored by what it did for that customer, every branch annotated with which way it went.

You can:

• Watch a run live as a customer interacts with their card, with warnings and errors highlighted as they happen.

• Inspect every conditional branch and toggle between the variable name (%{customerSegment}) and the value it resolved to for this customer (premium).

• See the analytic events that fired during the run, grouped against the step that produced them.

• Investigate step failures with the error message attached to the step that broke, alongside the events that surrounded it.

If you have a flowInstanceId from an analytics event or an API response and don't know which flow it belongs to, paste it into the Action Flow instance lookup tool under Tools and you go straight to the run.

Imagine you are running...

A retail bank. A high-value customer says the loan-approval card never showed up. You open the Action Flow's Runs tab, find their run, and within seconds see that an upstream Branch Flow routed them down the wrong path because their employment status was stale. You fix the data source and tell the customer it is on its way.

A telco's care team. You launched a new service-disruption flow yesterday, and you want to know it is actually doing the right thing for live customers, not just your test users. You watch a run in real time as a customer dismisses the first card, triggers the follow-up, and completes the task.

A digital bank rolling out a new branch condition. You ship a new conditional path on Friday. On Monday morning, you open the Runs tab for ten recent customers and visually confirm they routed correctly. Your QA loop just shrank from a sprint to a coffee break.

Why Atomic

Most engagement platforms, CDPs, and push providers stop at aggregate analytics: opens, clicks, conversion rates across cohorts. They cannot tell you what happened to one named customer on one flow at one moment, and that is exactly the question you ask when something looks wrong. Atomic's per-instance Runs tab is built around that question. You see the same graph you authored, with the actual run painted onto it.

Try it

The Action Flow Runs tab is in preview for selected organizations. To see what is possible, read the Action Flow Runs tab guide, or talk to your CSM about getting it turned on for your organization.