AI Dashboard
If you offer AI features to your customers, the AI Dashboard gives you platform-wide visibility into how much AI is being used, what it's costing, and whether the AI infrastructure is healthy.
Location: Admin → AI (/admin/ai)
📸 Screenshot: The AI Dashboard with usage stats, the model breakdown and the daily token chart.
Usage statistics (last 30 days)
| Metric | Meaning |
|---|---|
| Total tokens | Total AI tokens consumed across all customers. |
| Total runs | Number of AI generations/replies. |
| Error runs | Failed AI calls (helps spot misconfigured keys). |
| Average latency | How long AI responses take, in milliseconds. |
A daily token usage chart (last 14 days) shows the trend at a glance, and a top models list shows which models are used most by run count and token volume.
Provider distribution
See how many workspaces are using each AI provider (OpenAI, Claude, Gemini), and how many workspaces have at least one AI provider configured. This helps you understand adoption and plan capacity.
Vector database status
If you've connected Qdrant for semantic search, the dashboard shows whether it's configured, reachable and healthy. If Qdrant isn't configured, WhatsMine uses its built-in in-database fallback — so AI knowledge bases still work.
Knowledge base & chatbot stats
- Knowledge bases — total count and document statuses (indexed, processing, failed).
- Chatbots — total and how many are active.
Why this matters
- Cost control — AI usage drives your provider bill. Watch the token chart and the top spenders so there are no surprises.
- Billing — usage maps to plan limits (
ai_tokens_per_month), so you can price AI fairly. See Plans & Billing. - Reliability — a spike in error runs usually means an expired or invalid API key in Integrations.
➡️ Next: keep the platform running smoothly with System Tools.