> ## Documentation Index
> Fetch the complete documentation index at: https://docs.assistable.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Use the AI Replying indicator

> Interpret an in-progress GHL chat tag without mistaking it for message delivery.

For automatic GHL chat, Assistable attempts to add **AI Replying** to the contact before queueing response generation. The tag has a space; it is not `ai_replying`.

The run attempts to remove the tag when it finishes or is cleaned up, including some skipped and failed paths. These tag updates depend on GHL accepting the request, so the indicator can be delayed or remain after a failure.

## What it tells you

| Observation | What to check next |
| - | - |
| Tag appears | A reply run is being queued or processed; wait for its outcome |
| Tag disappears | Inspect the conversation to see whether a reply was sent |
| Tag remains unusually long | Check the latest messages and errors before taking action |
| Tag never appears | Check routing, pause settings, and whether GHL tag updates succeeded |

Neither the presence nor absence of this tag proves delivery. The strongest check is the actual outgoing message and its status in the customer conversation.

## Build an escalation workflow

If you monitor this tag in GHL, allow for the assistant's **Wait time**, normal processing, and provider delay. After the waiting period, check the latest message and error state before notifying your team.

Prefer a review or notification step before an automatic fallback message. If the AI reply arrives late, an unconditional fallback can send the customer two answers.

## When asking for help

Include the contact, conversation, approximate time, matching active tag, and whether the customer received a reply. Say whether the indicator remained or disappeared; do not use that observation as the only evidence of the failure.

This indicator applies to the GHL chat path. Direct conversations and voice calls do not use it as the same reply marker.

See [GHL chat setup](/v3/guides/chat/activate-ghl-chat) and [pausing a contact](/v3/guides/chat/pause-contact).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.