AI Dashboard
User Guide — A Single View of AI Activity, Cost, and Health
Module: AI Module › Configuration | Last updated: July 2026
Contents
- Overview
- Finding Your Way Around
- The Summary Cards
- Recent AI Activity
- System Health
- Mapping AI Acceptance
- Quick Reference
1. Overview
The Dashboard is the at-a-glance home for the AI Module. It gathers the numbers that matter most — how much AI is set up, what it's costing, what it's been doing lately, and whether everything is running smoothly — so you can check on AI in OnCoor without opening each screen one by one.
Everything on the Dashboard is read-only: it's a summary you look at, not a place you change settings. When something needs attention, it points you to the screen where you can act.
What you can see here
- How much AI is switched on — domains, models, and providers.
- What AI is costing this month.
- The most recent AI activity across your domains.
- Whether background AI work is healthy.
- How well the AI's mapping suggestions are landing.
WHERE TO FIND IT — The Dashboard is under Admin → AI → Configuration → Dashboard.
2. Finding Your Way Around
The Dashboard is arranged in four areas, top to bottom:
- Summary cards — the key totals across the top.
- Recent AI Activity — a running list of the latest requests (left).
- System Health — a set of status tiles (right).
- Mapping AI Acceptance — how useful the AI's mapping suggestions have been (bottom).
Each area is covered below.
3. The Summary Cards
Four cards across the top give you the headline numbers.
| Card | What it shows |
|---|---|
| AI Enabled Domains | How many data domains have AI switched on. |
| Active Models | How many AI models are switched on, across all providers. |
| Active Providers | How many AI services are switched on. It also flags any that are pending review. |
| Est. Monthly Cost | The estimated cost of AI usage so far this calendar month, based on the amount of text (tokens) used. |
LOCAL AI KEEPS THE COST AT ZERO — If you're running on the built-in local AI, the estimated cost stays at or near zero, because local models don't charge per use.
4. Recent AI Activity
This panel lists the most recent AI requests, newest first, so you can see what the AI has been doing at a glance.
| Column | What it shows |
|---|---|
| Date | When the request happened (shown as "Today", "Yesterday", or a date). |
| Domain | Which data domain it belonged to. |
| Activity | The kind of request — for example, a chat answer or a matching job. |
| PII | A marker if the request involved personal information. |
| Status | The outcome — Success, Failed, Partial, or Blocked. |
| Model | The model that handled it. |
Click VIEW ALL to open AI Observability, where you can see the full history, filter it, and open the detail of any request.
A QUICK VIEW ONLY — The activity list shows a short, safe summary with personal information hidden. For the exact prompts and responses, use Observability.
5. System Health
These tiles tell you whether the AI Module's background work is running smoothly. Each tile is calm when all is well and turns amber or red when something needs a look. A timestamp shows when the health was last checked.
| Indicator | What it tells you |
|---|---|
| Embedding Queue (pending) | How much record-matching work is waiting to be processed, and how much is in progress. A large backlog turns amber. |
| Stuck Workers | Whether any background jobs have stalled. Anything above zero turns red and needs attention. |
| In-flight RAG | How many chat requests are being worked on right now. |
| 24h Errors | How many AI requests failed in the last 24 hours (out of the total), including any that were blocked by a privacy or usage-limit rule. |
| Log storage health | Whether the behind-the-scenes log housekeeping is up to date. A warning here is one to raise with your administrator. |
THE DASHBOARD FLAGS, IT DOESN'T FIX — When a tile turns amber or red, it's pointing you to something to check. Stuck workers can be restarted from the Knowledge → Embedding Queue screen; most other warnings are best handled with your system administrator.
6. Mapping AI Acceptance
If you use AI mapping suggestions, this panel shows how helpful they've been over the last 7 days. The big number is the acceptance rate — how often people actively chose the AI's top suggestion. Below it, a breakdown shows what happened with each suggestion:
| Result | What it means |
|---|---|
| Accepted (top pick) | The user took the AI's first suggestion. |
| Modified (lower rank) | The user picked a different suggestion than the AI's top one. |
| Rejected (kept matcher) | The user kept the system's own match instead of the AI's suggestion. |
A higher acceptance rate means the AI's suggestions are landing well. If you don't use mapping suggestions yet, this panel simply shows that there's nothing to report.
7. Quick Reference
A fast lookup for the most common actions.
| I want to… | Do this |
|---|---|
| Get an overall picture of AI | Admin → AI → Configuration → Dashboard |
| See how much AI is switched on | Read the summary cards at the top |
| Check this month's estimated cost | Look at the Est. Monthly Cost card |
| See what the AI did recently | Read the Recent AI Activity list |
| Open the full activity history | Recent AI Activity → VIEW ALL |
| Check that AI is running smoothly | Scan the System Health tiles |
| See how useful mapping suggestions are | Read the Mapping AI Acceptance panel |
Source: OnCoor AI Module product documentation, written for end users. For the latest screens and options, always refer to the in-app interface.