Managing Conversations
Monitor, search, and manage customer conversations with your AI agent
The Conversations section gives you full visibility into every interaction between your customers and your AI agent. Use it to monitor quality, identify knowledge gaps, and step in when needed.
Conversation Statuses
Every conversation has a status that reflects where it is in its lifecycle:
| Status | Meaning |
|---|---|
open | Active conversation — the AI agent is responding normally |
awaiting_agent | Customer is waiting; the AI has flagged this for human review or the customer explicitly requested human help |
human | A human team member has taken over the conversation. The AI agent will not respond while a conversation is in this status |
closed | Conversation is resolved and archived |
Changing conversation status
You can change the status from the conversation detail view using the status dropdown at the top of the thread. Common workflows:
- Open → Human: take over a conversation from the AI to respond personally
- Human → Open: hand control back to the AI agent
- Any → Closed: mark a conversation as resolved
Viewing Conversations
Navigate to your agent's Conversations page to see a list of all chat sessions. Each conversation shows:
- Source — where the conversation originated (chat widget, Slack, Discord, etc.)
- Preview — the first few messages exchanged
- Timestamp — when the conversation started
- Status — current status (open, awaiting_agent, human, closed)
Click on any conversation to view the full message history.
Filtering and Searching
Use the toolbar at the top of the conversations list to narrow down what you're looking at:
- Search — find conversations by keyword across all messages
- Source filter — show only conversations from a specific channel (e.g. only Slack messages)
- Date range — focus on a specific time period
This is especially useful when investigating a customer issue or reviewing how your agent handles a specific topic.
Sentiment Analysis
FutureBase automatically scores the sentiment of each message as conversations progress. The sentiment score (positive, neutral, or negative) is visible in the conversation detail view and in the conversation list as a color indicator.
Use sentiment to:
- Quickly identify frustrated customers who need immediate attention
- Filter conversations by negative sentiment to prioritize review
- Track sentiment trends over time in your analytics
Sentiment is scored by AI on each incoming message and is recalculated as the conversation continues.
Understanding AI Responses
Each AI response is generated based on:
- Context — the most relevant content from your knowledge base
- Conversation history — previous messages in the same thread for continuity
- Agent configuration — any custom instructions or tone settings you've applied
If an AI response seems off, check whether the relevant information exists in your knowledge base. Often, poor responses are a sign of missing or outdated content.
Escalation and Handoff
When your AI agent encounters a question it can't confidently answer, or when a customer explicitly asks for human help, the conversation can be escalated.
Automatic Escalation
Your agent is configured to recognize when it should suggest human assistance. Common triggers include:
- Questions outside the scope of your knowledge base
- Customer frustration or repeated questions
- Explicit requests like "talk to a human" or "connect me to support"
Creating Tickets
From any conversation, you can create a Ticket to track the issue and assign it to a team member. Tickets integrate with your existing workflow via integrations like Linear, Jira, or the built-in ticket board.
Conversation Analytics
Over time, your conversations become a valuable data source. Look for patterns like:
- Common questions — if the same question comes up frequently, add it as an FAQ
- Unanswered topics — if your agent often can't find relevant content, fill those gaps in your knowledge base
- Peak hours — understand when your customers need the most support
Conversations are retained according to your plan's data retention policy. Check your Settings page for details.
Multi-Channel Conversations
If you've connected integrations like Slack, Discord, Telegram, or WhatsApp, conversations from those channels appear alongside your chat widget conversations in the same unified inbox. The Source filter helps you separate them when needed.
Each channel has its own nuances (e.g. thread support in Slack, group chats in Telegram), but all conversations are searchable and manageable from the same interface.