CORAA
University · AI governance

AI Use Log Register Generator.

Paste a record of where your team used AI tools: date, tool, task, class of data, who reviewed the output and what happened to it. The tool builds a tidy register you can print or take to Excel, and shows the entries that have no reviewer or no outcome.

Log entries

Paste from Excel or CSV. Use the data class names your firm uses. Do not paste client data itself, only a description of what the tool was used for.

This is a record-keeping aid. The AI policy of your firm, client confidentiality terms and data protection duties decide what may be used and by whom. Check the current rules.

Register summary
Entries6
Entries with no reviewer2
Entries with no outcome1
Entries needing follow-up3
ToolEntries
Drafting assistant2
Spreadsheet assistant2
Transcription tool1
Research assistant1
Data classEntries
Client financial2
Public2
Confidential1
Personal data1
RowToolData classFollow-up
3Spreadsheet assistantClient financialNo reviewer recorded on a sensitive data class
6Spreadsheet assistantClient financialNo outcome recorded
7Research assistantPublicNo reviewer recorded

A log shows that AI use is supervised.

Firms are being asked by clients, peer reviewers and their own quality systems how AI tools are used and who checks the output. A simple log answers that. It records the tool, the task, the class of data involved, who reviewed the result and what was done with it. Entries with sensitive data and no reviewer are the ones to fix first.

AI audit tool evaluation checklistAI governance workpaper

How the register is built

Each pasted line is one use of an AI tool. The tool reads the date, tool name, task, data class, reviewer and outcome by header name.

You list the data classes that need a reviewer. Entries in those classes with no reviewer are marked, and any entry with no outcome or no data class is marked too.

Counts are shown by tool and by data class. You can print the log with a sign-off line or download the Excel file.

Worked example — a missing reviewer

A spreadsheet assistant was used to draft commentary on client financial data and nobody is named as reviewer.

Inputs
Data classClient financial
ReviewerBlank
Classes needing reviewConfidential, Client financial, Personal data
Output
Follow-up flagNo reviewer recorded on a sensitive data class
The entry goes on the follow-up list so someone can review the output and record who did it.

Common mistakes

Pasting client data into the log
Describe the task. Do not copy the client information into the register.
Vague data classes
Agree a short list of class names so the counts mean something.
Logging after the fact
Entries made weeks later miss detail. Record the use when it happens.

Frequently asked questions

Is a log required?+
It depends on your firm policy, client terms and the rules that apply to you. Check the current requirements.
What columns are needed?+
A header row, then date, AI tool, task, data class, reviewer and outcome.
Is my data uploaded?+
No. The register is built in your browser.

Authoritative sources

ICAI — Standard on Quality Management (SQM 1) — Firm level quality management, including resources and technology. Check the current text.
Always confirm against the latest version of the source. Regulations evolve and amendments are common.
Related calculators
AI audit tool evaluation checklist →AI governance workpaper →SQM 1 gap assessment →
Share this tool
Last reviewed: 2026-10-17 · For informational purposes only — not professional advice.