# AI assistance activity

Review AI assistance activity and understand how token usage, cost estimates and execution limits relate.

*Release 3.3 · admin, supervisor*

**Settings › AI › Features** · **Settings › AI › Accounts**

## What it is

INSIDE records AI activity and available model usage measurements. **AI Usage** report columns show how operators use AI assistance, including whether they use, edit or ignore suggestions. Generation records can also contain token counts, duration and an estimated cost when model pricing is available.

## Who uses it and why

- Supervisors review AI assistance adoption alongside operator activity.
- Administrators tune feature limits and ask Powerfront to review account limits or usage estimates.

## Where to find it

Look for the **AI Usage** columns in reports that include AI assistance activity. Feature limits are in **Settings › AI › Features**.

Powerfront manages **Settings › AI › Accounts**. Ask Powerfront to change account function limits or help investigate detailed token usage and cost estimates.

## How it works

For AI generation, INSIDE records available input, output and total token counts and the time taken. A conversation can involve several generations, tool calls and separate summary work, so one response is not necessarily one AI request.

For operator assistance, the report distinguishes **Suggestions**, **Used**, **Edited**, **Ignored**, **Resuggestions** and **Minimized**. **Ignored** means the operator sent a message directly after a suggestion. **Edited** counts a suggestion that the operator edited and then used. Other columns include **Fix**, **Re-Phrase**, **Ask AI** and **Translations**.

Estimated generation costs use available model pricing and are recorded in US dollars. Missing pricing or measurements can leave the estimate unavailable.

## Configure it

With the **Settings** role and the **AI** permission, or the **Developer** role, open the feature in **Settings › AI › Features** and review **Max Tokens**. Leave it at zero to inherit the account default.

Ask Powerfront to change **Max Function Count** or **Max Function Seconds** on the account. These default to 3 and 30 seconds. Tell Powerfront which limits suit the work the feature needs to do.

### Settings

| Setting | Where | Applies to | Default | Who can change it | What it does |
|---|---|---|---|---|---|
| Max Tokens | Settings › AI › Features | website | 0, use the account default | Settings role, Developer role | A positive value overrides the feature's inherited token limit; it is not a monetary budget. |
| Max Function Count | Settings › AI › Accounts | website | 3 | Powerfront, on request | Bounds repeated AI function work before the flow finishes without further tools. |
| Max Function Seconds | Settings › AI › Accounts | website | 30 seconds | Powerfront, on request | Bounds the time available for repeated AI function work before a final response. |

### Permissions

| Permission | Lets you |
|---|---|
| Settings | Change feature token limits when AI is allowed under Settings Permissions. |
| AI | Allows AI configuration within Settings Permissions. |
| Developer | Change AI feature configuration. |

## Limits and good to know

- Estimated cost is not an invoice amount. Usage counts and money estimates measure different things.
- A token limit or function limit is not a monthly spending cap.
- Reaching a flow budget can disable further tools while still allowing a final response. Do not treat it as a guarantee of no additional token use.
- Context compaction and citation summaries can create additional usage.
- A missing usage entry or estimate does not prove that no AI work occurred.

## Depends on and works with

### This feature uses

- **Configured in** [Roles and permissions](ADM-008.md): Roles control access to feature configuration and reporting.
- **Configured in** [AI feature settings and change history](AI-001.md): Feature configuration sets the feature token override.
- **Configured in** [AI run limits](AI-014.md): Execution limits bound repeated generation and tool work.
- **Configured in** [AI accounts and shared defaults](AIP-001.md): Powerfront manages account defaults and function work limits.
- **Works better with** [Copilot reply suggestions](AI-006.md): Copilot interactions supply suggestion adoption signals.
- **Works better with** [Context Compiler for AI conversations](AI-015.md): Context compaction can add separate AI usage.
- **Works better with** [AI answer citations and summaries](AI-028.md): Citation summaries can add separate AI usage.
- **Works better with** [AI model catalogue](AIP-002.md): Known model pricing allows estimated cost calculation.
- **Works better with** [AI usage and estimated cost](AIP-023.md): Generation usage records add token counts and cost estimates to AI activity analysis.
- **Related** [AI assistance, translation and sentiment reports](REP-026.md): AI assistance reports show operator activity and suggestion-adoption counts.

### Used by

- **Related** [Copilot reply suggestions](AI-006.md): AI Usage records whether operators use, edit or ignore Copilot suggestions.
- **Related** [AI usage and estimated cost](AIP-023.md): Compare token usage with AI assistance activity when reviewing resource consumption and operator adoption.

## FAQ and troubleshooting

### Why do I see AI activity but no estimated cost?

The model price or the required usage measurements may be unavailable. Ask Powerfront to check the account and model configuration.

### Are Used and Edited the same?

No. **Used** records use of a suggestion; **Edited** records editing a suggestion before using it. Review both when assessing adoption.

### Will Max Tokens stop monthly AI spending at a fixed amount?

No. It is an execution setting, not a monetary quota. Discuss account usage and limits with Powerfront.

### Why can a short answer use many tokens?

Usage can include the conversation context, tool results and multiple generations as well as the visible answer. Separate summarisation work can add further usage.