> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getthread.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Super Magic Usage Dashboard
> See how technicians and external AI agents use Thread's AI — tool calls, write actions, weekly pace, per-technician adoption, skill usage, automation share, and Thread MCP activity.
The Super Magic usage dashboard shows how your technicians — and any external AI agents connected through Thread MCP — are actually using Thread's AI capabilities. Use it to prove adoption, find the technicians who haven't started, and see which skills are running versus sitting idle.
Open it from **Thread admin → General → Analytics**.
## What this dashboard answers
* Is AI adoption **growing week over week**, or did it plateau after launch?
* Which **technicians** are using it most, and which haven't started?
* Which **tools** are being used, and how much of that is **reading** versus **writing**?
* Which **skills** exist, and which are actually being run?
* How much activity is **technician-driven** versus **fully automated**?
* What are **external AI agents** doing through Thread MCP?
## Filters and time windows
A tab bar at the top switches between views, and the selected tab is reflected in the page URL — so you can link someone straight to a specific view.
| Filter | Where it applies |
| ----------------------------------- | ----------------------------------- |
| Last 30 days, versus the prior 30 | Overview KPI cards and trend arrows |
| 12-week window | The Pace tab |
| 7, 14, 30, 60, 90 days, or all time | The Technicians tab date picker |
| Search, category, read/write type | The Tools tab leaderboard |
| Search and sort | The Skills tab card grid |
## What's on the dashboard
### Overview tab
Ten KPI cards covering the last 30 days — tool calls, sessions with actions, write actions, actions per session, tickets touched, distinct tools used, active technicians, active skills, interactive actions, and automation action share. Each carries a tooltip explaining exactly what it measures, and most show a trend arrow against the prior 30 days.
Alongside them: a **Magic insights** narrative summarizing adoption trends in the style of a quarterly business review, a weekly activity trend chart, a donut of actions by category, and tables of top technicians and top tools.
### Pace tab
Week-by-week charts across a 12-week window for tool calls, sessions, write actions, and distinct tools used, plus a stacked chart showing the category mix by week and a detailed weekly table. This is the tab that tells you whether adoption is compounding.
### Technicians tab
A searchable, sortable technician table with a date-range picker, plus KPI cards and charts ranking technicians two ways: by **volume** and by **depth of use** — actions per session. The second ranking is the more interesting one; it finds the technicians who have genuinely changed how they work.
### Tools tab
Breakdowns of actions by category and by read-versus-write split, side-by-side comparisons of read versus write and technician-driven versus automated activity, and a full tool leaderboard you can search by name or filter by category or read/write type.
### Skills tab
Covers your library of reusable prompts. KPIs for total, personal, and shared skills; a scope breakdown chart; a cumulative creation-over-time chart; and a searchable, sortable card grid. Clicking a card opens the full skill description and instructions, with a copy button.
### Automation tab
Separates technician-driven **interactive** activity from fully automated **flow run** activity, with KPIs, a weekly comparison chart, an action-mix breakdown by category, and a skill execution leaderboard showing which skills are actually running versus configured but idle.
### Thread MCP tab
Covers usage by **external** AI agents — Claude, Cursor, ChatGPT — calling into Thread's tools. KPIs for successful and failed calls, success rate, active members, partners, and workspaces, and call duration, plus an AI-written insights summary, weekly trend and client-app charts, and leaderboards for tools, technicians, and partners.
## Metric definitions
| Metric | Definition |
| ----------------------- | ------------------------------------------------------------------------------------------------------ |
| Tool call | One invocation of a Thread tool by Super Magic or an external agent. |
| Session with actions | A conversation in which at least one tool call was made — as opposed to one that only asked questions. |
| Write action | A tool call that changed something: a ticket update, a note, a time entry. |
| Actions per session | Write and read actions divided by sessions — a measure of depth, not reach. |
| Tickets touched | Distinct tickets that had at least one AI action taken on them. |
| Distinct tools used | How much of the available tool surface your team has actually reached for. |
| Active technicians | Technicians with at least one action in the window. |
| Active skills | Skills that were actually run in the window, as opposed to created. |
| Interactive actions | Actions a technician started in a conversation. |
| Automation action share | The share of all actions that came from a Flow rather than a person. |
| Flow run | An action taken by a Super Magic Agent fired from a Flow, with no one confirming each step. |
## Things to watch
**AI-written summaries on the Overview and Thread MCP tabs are drafts.** Both carry a review-before-sharing-externally disclaimer for a reason — read the wording before it goes to a client or a board.
* **Volume and depth tell different stories.** A technician with high tool calls may be running one shallow action repeatedly; actions per session is the better adoption signal.
* **Total skills is a vanity number; active skills is the real one.** The Automation tab's execution leaderboard is where you find skills that were configured and then forgotten.
* **A rising automation share isn't automatically good.** It means Flows are doing more of the work — which is the goal, provided the interactive number isn't falling because technicians stopped using Super Magic.
* **Thread MCP activity is separate from in-Thread usage.** It measures external agents calling Thread's tools, so it won't reconcile with the Overview tab's technician numbers.
* **Data refreshes daily**, so today's activity generally appears tomorrow.
## Next steps
Drive the adoption these numbers measure.
Find shared skills worth putting in front of your team.
## Related topics
- [Service Team Performance & AI Usage Dashboard](/analytics/service-team-performance-dashboard.md)
- [Getting Started with Magic Analytics](/analytics/getting-started-with-magic-analytics.md)
- [Voice AI Summary Dashboard](/analytics/voice-ai-dashboard.md)