---
title: "Explain current stage, recorded reach and stage transitions"
description: "Understand stage populations and recorded movement without mistaking incomplete observations for a conversion funnel."
language: en
canonical: https://mrailabs.com/docs/en/conversations-and-customers/conversation-stage-analytics/
published: 2026-10-27T15:51:31.099097
last_verified: 2026-10-27T15:51:31.099097
audience: "For customer operations teams"
availability: "Beta feature"
---

# Explain current stage, recorded reach and stage transitions
## Why do stage counts differ in Mr. AI?

Mr. AI separates Currently here, a snapshot of eligible active customers now, from Recorded reach in period, distinct customers observed in a stage during the selected dates. A customer may appear in several reached stages. These totals describe different populations and are not a conversion funnel.

[Open this task in Mr. AI](https://access.mrailabs.com/client/actions/performance-report)

![Current stage snapshots and recorded stage reach](https://mrailabs.com/api/docs/assets/dashboard-images/help-20261019-stages-en.png?v=219addec1cf8)

*Current stage snapshots and recorded stage reach. Actual interface with synthetic examples; names, counts and prices are illustrative. Availability can vary by account. Select the image to enlarge it.*

1. Open Performance Analytics and review the selected dates.
2. Compare Currently here with Recorded reach in period.
3. Switch to transition arrivals when you want arrivals that exclude initial observations.
4. Choose a stage, then Incoming, Outgoing or Both directions to inspect recorded direct moves.
5. Read transition events separately from distinct customers and check saved stage names and Builder versions.
6. Use directory links to inspect customers, then recheck that directory’s own filters and population.

## Read the evidence limits

Recorded reach includes initial observations and can include archived customers observed in the period. Current position excludes dates. Incoming and outgoing traffic count transition events, not unique people. Repeated occurrences do not prove a customer loop, and a missing event does not prove the customer never visited a stage.

## Common questions

### Why does the directory count differ?

The customer directory uses its own filters for active customers, visitors and channels. Ever reached can include the current stage when earlier history is incomplete. Recheck the directory filters; the Analytics dates are not carried across.

### Can I add stage totals to count customers?

No. One customer can reach multiple stages.

### Does an unknown event type prove conversion?

No. Review the evidence warning; an unknown record must not be interpreted as a confirmed transition.

[Filter, review and export the customer directory](https://mrailabs.com/docs/en/conversations-and-customers/manage-customer-profiles-and-contact-data/)

[Design the conversation journey](https://mrailabs.com/docs/en/knowledge-and-behavior/conversation-builder/)

## Check the result

Separate current stage, recorded reach and transitions. Review the period and evidence before comparing stages.

## Related guides

- [Mr. AI v2.5.0 — Reports that connect conversations to business decisions](https://mrailabs.com/docs/en/product-updates/release-2-5-0/): Review conversation activity, customer progress and saved AI reports to understand what customers need and decide what to improve next.
- [Generate, review and print an AI business report](https://mrailabs.com/docs/en/conversations-and-customers/ai-business-reports/): Choose a report question, review its time scope and points quote, then use saved evidence and recommendations responsibly.
- [Organize customer conversations with reusable labels](https://mrailabs.com/docs/en/conversations-and-customers/conversation-labels/): Create team-defined labels, assign several labels to a customer, and filter Chats without changing journey stages.
- [View customer images and play voice messages](https://mrailabs.com/docs/en/conversations-and-customers/customer-images-and-voice-messages/): Inspect original customer images and recordings, distinguish AI transcription, and recover from an unavailable attachment.