---
url: /relyai-acellemail/features/observability.md
description: >-
  See and audit every AI action — a usage dashboard, full conversation replay,
  tool-call forensics with CSV export, feedback and sentiment, a self-improve
  view, and a read-only permissions matrix.
---

# Observability

Every AI request, conversation and tool call is logged. Each surface below is its own entry in the **RelyAI** sidebar group.

::: info Logging never stops
Turning a surface's feature flag off on **RelyAI → Settings → Feature Flags** hides the *view* only. The underlying logging keeps running.
:::

## Usage Dashboard

**RelyAI → Usage Dashboard** — the health-and-cost picture at a glance.

| KPI | Shows |
| --- | --- |
| **Total runs** | Across your AI activity, and how many task keys they span |
| **Success rate** | With the success/failure split |
| **Average duration** | End-to-end, in milliseconds |
| **Feedback** | The 👍/👎 tally |

Additional dashboard breakdowns include:

* **By-engine table:** Run counts, success rates, token usage (in/out), and latency statistics.
* **Dedicated subtabs:** Task runs, agent actions, feedback summaries, overview KPI strip, and customer/user rollups.

![The RelyAI usage dashboard](/relyai-acellemail/img/relyai-observability-usage.png)
*RelyAI → Usage Dashboard*

::: info Where P95 latency lives
P95 latency is on the **Audit Viewer**, not the Usage Dashboard.
:::

## Conversations

**RelyAI → Conversations** lists every multi-turn session. Drill in to **replay the messages, tool calls and feedback** exactly as they happened.

## Audit Viewer

**RelyAI → Audit Viewer** — tool-call forensics: every agent action, with **filters**, a **status time-series**, **top failures**, **P95 latency**, and a dedicated **destructive-ops filter**. Export any filtered view to **CSV**.

## Feedback

**RelyAI → Feedback** collects **up/down ratings** from every AI surface, alongside **sentiment** categorisation.

## Self-Improve

**RelyAI → Self-Improve** surfaces optimization targets based on operational data:

* **Failure clustering:** Groups tool errors by tool, error class, and knowledge domain across the last 7 days.
* **Quality tracking:** Monitors flagged hallucinations and sentiment trends.
* **Knowledge hot zones:** Highlights common support queries where documentation is lacking.

## Permissions matrix

**RelyAI → Settings → Permissions** provides a **read-only matrix** of tool access:

* **Introspection only:** Reflects existing AcelleMail role permissions without altering them.
* **Transparency:** When enabled, customers can access **View my permissions** to review their access levels.

## Related pages

* [Agent tools](/relyai-acellemail/features/agent-tools) — the actions the audit viewer records
* [Feature flags and permissions](/relyai-acellemail/guides/feature-flags-and-permissions) — hide a surface without stopping logging