Supervisor Dashboard
The supervisor dashboard is the real-time operational view for everyone responsible for running a support team — queue depth, agent availability, active bot conversations, and pending chat offers, all in one place and updated live as things change. This guide covers what the dashboard shows, how to drill into specific teams and agents, the actions supervisors can take from the dashboard, and how to use it effectively during both busy periods and quiet ones.
Overview
The dashboard is designed for one job: giving supervisors enough information to make good operational decisions right now. That means:
- Real-time updates. Queue numbers, agent statuses, bot conversations, and pending offers all update live. No refresh button, no polling lag. When an agent goes available, the dashboard reflects it within a second.
- Multi-team aggregate by default. Supervisors who oversee more than one team see the combined view across all their teams, with per-team breakdowns one click away.
- Drill-down, not dashboard-hopping. Any metric can be expanded into its constituents — click the queue depth to see the actual waiting conversations, click agent availability to see who's on what, click bot conversations to see the active bot interactions.
- Actions in place. When something needs intervention, the supervisor can reassign, take over, or adjust routing from the dashboard without navigating to a separate admin interface.
Top-Level Metrics
The top of the dashboard shows the key numbers at a glance across all teams the supervisor oversees.
| Metric | What it means |
|---|---|
| Queue depth | Total conversations waiting for an available agent. Broken down into pending offers (currently being offered to an agent) and waiting (not yet offered). |
| Active conversations | Conversations currently being handled by an agent. Includes in-progress work and ticket conversations that haven't yet been resolved. |
| Available agents | Agents currently set to Available status and below their capacity limits. These are the agents who can receive new offers right now. |
| Bot conversations | Active AI bot conversations. Includes the count of bots currently in conversation with customers and a resolution-rate indicator for recent bot-handled conversations. |
| Oldest waiting | How long the longest-waiting conversation has been in queue. The single best leading indicator of customer experience degradation. |
Queue Breakdown
Queue depth as a single number is less useful than understanding its composition. The dashboard breaks queued conversations into:
- Pending offers — conversations currently being offered to an agent but not yet accepted. Ticks down as agents accept and ticks up as new conversations enter routing.
- Waiting for agent availability — conversations that finished flow processing but can't be offered because no agent has capacity. This is the "real" queue that needs agent availability to clear.
- In bot conversation — conversations currently being handled by an AI bot. These may or may not escalate to an agent.
- In flow — conversations in progress through a conversation flow that hasn't yet routed to a team.
Seeing the composition matters: a queue of 20 that's entirely pending offers is very different from a queue of 20 waiting for agents who are all at capacity. The first will clear quickly; the second needs intervention.
Agent View
The agent view shows every agent on the teams the supervisor oversees, grouped by team and with real-time status.
| Column | Shows |
|---|---|
| Agent | Name and display avatar. |
| Status | Current display label (e.g., "Available", "Lunch Break", "In Meeting"). Custom labels appear with their configured icons. |
| Functional status | Whether the agent is Available, Not Available, or On Ticket — the status the routing engine sees. |
| Active | Number of active conversations vs. the team's capacity limit (e.g., "3 of 5"). At-a-glance signal for who's busy. |
| Last active | Time since the agent's last customer message or status change. Quickly identifies agents who may need a check-in. |
Clicking an agent expands their row to show each active conversation, recent activity, and a quick-message link for supervisor coaching.
Bot Conversation Visibility
When AI bots are handling customer conversations, supervisors need visibility into what they're doing. The dashboard surfaces:
- Active bot conversations — each one shows the customer, the bot handling it, how long it's been running, and a short summary of the current state.
- Recent resolution rate — percentage of bot conversations in the last hour that resolved without escalating. Leading indicator of bot effectiveness.
- Escalations — bots that handed off to humans, with reason (customer request, out-of-scope, guardrail triggered, knowledge gap).
- Knowledge gaps — questions the bot couldn't find relevant knowledge base content for, aggregated over the recent window. Useful signal for the next KB content update.
Supervisor Actions
From the dashboard, supervisors can take several actions without navigating away:
Reassign a conversation
Click a conversation in any view (queued, active, bot) and reassign it to a different agent or team. Useful when one agent is overloaded, when a ticket needs a specialist, or when a bot is struggling with a specific case.
Take over a conversation
Supervisors can take over a conversation themselves when needed — for VIP escalations, for cases requiring a policy exception, or just to help during a spike. Supervisors are agents with more permissions; they can pick up any conversation their teams have access to.
Temporarily adjust routing
For short-term adjustments (a team member went home sick, an unexpected surge, a promotional campaign causing spikes), supervisors can toggle queue overflow rules or temporarily change agent selection strategies from the dashboard without full admin access.
Send team messages
Quick messages to all agents on a team or to specific agents — useful for announcing a spike, sharing a workaround, or coordinating during an incident. These post to the team's internal chat.
Saved Views and Filters
Supervisors can save dashboard configurations for different contexts: one view that shows only their primary team, another that shows VIP customer conversations across all teams, another that filters to just bot conversations approaching escalation.
Saved views are per-user and don't affect other supervisors. They're useful for quickly switching between operational contexts (morning triage, afternoon QA, incident response).
Best Practices
- Check Oldest Waiting regularly. This single number is the best indicator of how the customer experience feels right now. If it's growing, act.
- Drill into the queue composition. Don't just react to queue depth; understand why. Pending offers ticking up suggests agents are slow to accept; waiting conversations piling up suggests you're at capacity.
- Use bot visibility proactively. Monitor bot conversations that are approaching escalation — you can sometimes catch issues that warrant a pre-emptive takeover or a knowledge base fix.
- Communicate spikes early. If you see queue depth rising, let the team know immediately through internal chat. Agents respond faster to a "we're getting slammed, heads up" message than to just watching numbers climb.
- Review bot knowledge gaps weekly. The questions bots couldn't answer are the content backlog. Feeding them back into the knowledge base improves bot quality over time.
Related Topics
- Team Management — configuring the teams, routing strategies, and agent statuses the dashboard reflects.
- AI Bot Configuration — configuring the bots whose conversations appear on the dashboard.
- Reporting & Dashboards — historical and aggregate reporting beyond the real-time operational view.