Skip to main content
The Apps dashboard gives you a unified view of how your conversational workflows are performing — from the moment a user starts an interaction to the point they complete (or abandon) a booking or service request. Use this dashboard to pinpoint friction in your WhatsApp flows, compare the relative adoption of your automation workflows, and keep track of which app versions are deployed in production.

​
WhatsApp Flow Funnel

The WhatsApp Flow Funnel visualizes how users progress through the key steps of your appointment booking journey. Each stage in the funnel represents a distinct interaction point, and the count at each step tells you how many users reached that stage during the selected time period. The funnel tracks these steps in order:
1

Select Location

The user sees and interacts with the location selector. This is typically the first step after the initial greeting and intent capture. A high count here indicates strong top-of-funnel engagement.
2

Select Consultant

The user picks a specialist or doctor. Drop-off between Select Location and Select Consultant often signals that the available consultant list is too long, unclear, or doesn’t match the user’s expectation.
3

Available Slots

The user is shown open appointment times for their chosen consultant. A significant drop here can indicate slot availability issues or slow response times pulling calendar data.
4

Confirmation

The user confirms their appointment. This is your conversion step. Comparing confirmation volume against the top-of-funnel count gives you your end-to-end booking completion rate.
Calculate your funnel completion rate by dividing Confirmation count by Select Location count. Aim to improve this ratio over time by addressing the step with the steepest relative drop-off first.

​
Reading drop-off signals

When you notice a large gap between two consecutive funnel steps, investigate the specific workflow node responsible for that transition. Common causes include:
  • Ambiguous button labels or response options at that step
  • A backend lookup (e.g., fetching consultant availability) that times out or returns an empty result
  • A mismatch between the user’s stated intent and the options your workflow presents
Funnel data reflects users who actively interacted with that step — users who abandoned the conversation before reaching a step are not counted in that step’s total.

​
Workflow Execution Metrics

The Workflow Execution Metrics section displays a bar chart showing how many times each of your core automation workflows was triggered during the selected date range. This helps you understand which workflows are doing the heaviest lifting and which may be underutilized.

Live RAG Flow

Counts every time the Retrieval-Augmented Generation pipeline is invoked to answer a user query from your knowledge base. A high count here indicates strong question-answering demand from your users.

Call Processing

Tracks the number of voice interactions that were routed through the call processing workflow, including speech-to-text transcription and intent routing. Compare this against your transcription count in CX Ops to measure end-to-end voice pipeline health.

Practo Booking Flow

Records how often the Practo-integrated appointment booking workflow was triggered. Cross-reference this with your WhatsApp funnel’s Confirmation count to verify that triggered bookings are completing successfully.

​
Using execution counts to improve automation

If a workflow has a very low execution count compared to others, determine whether users are failing to reach the entry point for that workflow or whether the workflow itself is not being triggered correctly. Check your intent classification rules in the Ontology dashboard to confirm that the right topics map to the right workflow entry points.
A sudden spike in a workflow’s execution count can indicate either a surge in genuine user demand or a routing error that’s sending traffic to the wrong workflow. Compare the execution count trend against your Trending Topics data in CX Ops to confirm whether the spike reflects real intent.
If one workflow is consistently handling far more traffic than others, review whether some of that load should be distributed — for example, by expanding the knowledge content available to the Live RAG Flow so that common queries don’t always fall through to a more expensive booking workflow.
Set a consistent date range (for example, the last 30 days) when comparing workflow execution counts across reporting periods so your trend analysis stays accurate.

​
App List and Version Tracking

The App List section shows every application registered in your Neuroncx workspace, along with its current version number. Each entry corresponds to a deployable unit — such as a specific WhatsApp bot configuration or a RAG-enabled assistant — that runs one or more workflows.
Version numbers increment automatically each time you publish a change to an app’s configuration. Keeping track of versions lets you correlate performance changes in your funnel or workflow metrics with specific deployments.

​
Best practices for version management

  • Label releases consistently — before publishing a new version, document what changed (e.g., updated consultant list, revised confirmation message, new slot-selection logic) so you can trace any metric shifts back to a specific change.
  • Compare metrics across versions — filter your funnel and workflow execution charts by date ranges that correspond to specific version windows to understand the impact of each release.
  • Roll back with confidence — if a new version causes a measurable drop in funnel completion or workflow execution quality, use the version history to identify the previous stable release and restore it promptly.
Deploying a new app version updates the live experience for all users on that channel immediately. Always validate changes in a staging environment before publishing to production.