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

# HighLevel Conversations

> Track AI agent conversations, outcomes, and lead attribution in real time.

The Conversations dashboard gives you visibility into every AI agent conversation happening in your HighLevel account — what's being said, how leads are being handled, and which conversations are converting.

Conversations are synced automatically every night. New conversations from HighLevel appear in WisePilot within 24 hours. After sync, conversations are scored for quality and outcomes are classified. For accounts with a large conversation history, the initial import may take a few daily sync cycles to complete — each cycle processes what it can and the next picks up where it left off. Outcome classification (detecting bookings, opportunity wins) always runs on every cycle. See [Automation & Data Freshness](/optimization/automation) for the full schedule.

Navigate to **CRM → Conversations** in the sidebar.

## Dashboard Overview

The main view shows a filterable list of all conversations with key data at a glance:

| Column       | Description                                               |
| ------------ | --------------------------------------------------------- |
| **Contact**  | Lead name and contact details                             |
| **Agent**    | Which AI agent handled the conversation                   |
| **Outcome**  | Booked, qualified, nurture, lost, or escalated            |
| **Messages** | Number of messages exchanged                              |
| **Duration** | How long the conversation lasted                          |
| **Source**   | Which content/CTA originated this lead (attribution data) |
| **Date**     | When the conversation occurred                            |

## Conversation Detail

Click any conversation to see the full detail view:

### Transcript

The complete message history between the AI agent and the contact. Each message shows:

* Sender (agent or contact)
* Timestamp
* Message content

### Outcome Classification

How the conversation ended:

| Outcome       | Meaning                             | Next Step                   |
| ------------- | ----------------------------------- | --------------------------- |
| **Booked**    | Appointment scheduled               | Follow up to confirm        |
| **Qualified** | Lead meets criteria but didn't book | Enter nurture sequence      |
| **Nurture**   | Interested but not ready            | Enter nurture sequence      |
| **Lost**      | Disengaged or disqualified          | Review for pattern insights |
| **Escalated** | Handed off to human agent           | Human takes over            |

### Attribution Data

If the lead came through a tracked CTA:

* **Content asset** — Which page they visited
* **CTA clicked** — Which call to action they engaged with
* **Attribution ID** — The `cr_attrib` that links this conversation to the content funnel

This tells you not just "a lead booked" but "a lead who read the pricing page and clicked 'Talk to Sales' booked."

### Contact Information

HighLevel contact details including:

* Name, email, phone
* Custom fields (including `cr_attrib`)
* Contact tags and pipeline stage
* Team member assignment

## Filtering

Narrow conversations using:

* **Agent** — Focus on a specific AI agent's conversations
* **Outcome** — Show only booked conversations, only lost, etc.
* **Date range** — This week, this month, custom range
* **Team member** — Filter by assigned team member
* **Source** — Filter by originating content asset or CTA

## Conversion Funnel

The dashboard includes a visual conversion funnel showing:

1. **Total conversations** initiated
2. **Engaged** — Multi-turn conversations (contact replied at least once)
3. **Qualified** — Conversations that reached qualification criteria
4. **Booked** — Conversations resulting in appointments

Each transition shows the drop-off percentage. A steep drop between "Engaged" and "Qualified" may indicate the agent isn't asking the right qualifying questions. A drop between "Qualified" and "Booked" suggests the booking CTA or scheduling flow needs work.

## Taking Action

| Finding                                  | Action                                                                                     |
| ---------------------------------------- | ------------------------------------------------------------------------------------------ |
| Low engagement rate                      | Review agent [prompt](/integrations/highlevel/agents) — is the opening message compelling? |
| High "lost" rate from a specific source  | The content/CTA may be attracting the wrong audience — adjust targeting                    |
| Conversations are long but don't convert | Agent may be looping. Simplify the prompt and add clearer decision paths.                  |
| One agent outperforms others             | Compare prompts and replicate the winner's patterns                                        |

For deeper analysis, see the [Conversation Intelligence](/reports/conversation-intelligence) and [Conversation Impact](/reports/conversation-impact) reports.