Skip to content

Checkpoint Node

Measure passage volume and response rate at specific points in your flow.

What is it?

The Checkpoint node is a non-blocking node you explicitly place in the flow to record analytics events. Unlike the previous model (which tried to infer follow-ups automatically from the session timeout), the Checkpoint gives you full control over where and what to measure.

The node lives in the 📊 Statistics category of the Flow Builder's side panel. When a conversation passes through it, the flow continues immediately to the next node — registration happens in the background without blocking the conversation.

What is it for?

  • Know how many conversations pass through a specific point in the flow (volume)
  • Measure whether the customer replies within a configurable window after passing that point
  • Capture a snapshot of flow variables at the moment of passage
  • Compare performance across different points within the same flow
  • Get a summarized indicator on the sales funnel without opening detailed reports

Configuring the node

  1. In the Flow Builder, open the 📊 Statistics category
  2. Drag the Checkpoint node onto the canvas
  3. Connect the input (left) and output (right) in the flow
  4. Click the node body to open settings:
OptionDescriptionDefault on create
Track passage (volume)Records each passage as a volume event (passed)Enabled
Track customer responseStarts the response window countdown; customer messages within the window mark as respondedEnabled
Response windowTime in minutes (fixed value) or numeric flow variable2880 min (48h)
Save flow variablesStores a snapshot of variables at the moment of passageDisabled

💡 Where to place it?

Place the Checkpoint after the node that sends the follow-up (or re-engagement) message and before the next branch in the flow. That way you measure exactly the impact of that message.

Response window

The window defines how long the system waits for the customer's reply after passing the checkpoint. It can be configured at three levels (in cascade):

  1. On the node itself — fixed value in minutes or numeric flow variable (recommended)
  2. On the flow — legacy followup_response_window_hours field (still accepted by the backend, but removed from the UI)
  3. On the organization — in Settings → Organization, Follow-up response window field (default: 48h)

If no value is configured, the system uses 48 hours (2880 minutes).

What counts as a "response"?

Any message sent by the customer after passing the Checkpoint, as long as it's within the configured response window. If the customer doesn't reply within the window, the event is marked as expired (no response).

StatusMeaning
Passed (passed)Volume recorded without tracking response (only "Track passage" enabled)
Waiting (waiting)Checkpoint with response tracking active; waiting for customer message
Responded (responded)Customer sent a message within the response window
Expired (expired)The response window ended without a reply from the customer

Viewing reports

Full flow report

  1. Open the flow in the Flow Builder
  2. Click Follow-up Analytics in the top bar
  3. The panel shows:
    • Recorded passages and follow-ups sent (checkpoints with response tracking)
    • Response rate and average response time
    • Evolution over time — chart with the rate over the period
    • By node — table with node name, sent, responded, expired, and rate

Single checkpoint report

  1. On the Checkpoint node itself, click View report for this point
  2. The panel opens already scoped to that node (the "By node" table is hidden)

Available filters

  • Start / end date — restricts the analyzed period
  • Service channel — filters by specific channel
  • Assigned agent — filters by agent assigned to the chat
  • Refresh — button to manually reload data

Sales Funnel indicator

In CRM → Sales Funnel, a summary card shows the organization-wide response rate with total events in the period.

Change from the previous model

In v2026.7.3, automatic tracking coupled to "Wait for Response" timeouts and AI Agent reschedules was removed. That mechanism produced incorrect records by conflating response timeouts with other internal flow processes.

As of this version, only explicitly placed Checkpoints generate metrics. If you were using automatic follow-ups and want to keep measuring effectiveness, add a Checkpoint after each re-engagement message.

Limitations and notes

  • The node must be published in the flow to record events in production
  • Changing the response window does not affect already recorded events, only new ones
  • Events recorded during flow testing also count toward metrics
  • Registration is best-effort: database failures never block the flow

Next Steps

Documentation constantly being updated