Why the data going into AI tools needs visible governance
Record what data enters each AI tool, including pasted text, files and inputs passed by embedded features. Classify the information and check the proposed use against the tool’s settings, terms and the firm’s approval rules.
For owner-led firms, professional services teams and regulated businesses, the answer affects confidentiality, data protection, vendor risk, audit evidence and client trust. The tool may look like a writing assistant, but the input can include client names, contract clauses, deal details, customer complaints, claims notes, board papers, financial data or internal operating procedures.
The practical starting point is not a long policy document. It is a clear view of data categories, where each category may be used, who can approve exceptions and how outputs are checked.
The safest approach is to classify inputs before teams use AI
Classify input data before approving an AI use case. Trace what enters the tool, where it is processed, whether it is retained, who reviews the output and what evidence the team keeps for that decision.
Use the four groups below as a starting point, then check them against the firm’s actual information and restrictions:
- Open information: public web copy, published policies, generic examples and non-confidential templates.
- Internal information: process notes, meeting summaries, draft plans and operational data that should not leave approved systems.
- Client or customer information: documents, names, facts, case details, financial records, claims data or support history.
- Restricted information: special category data, privileged legal material, credentials, security details, M&A plans, disciplinary records and anything covered by strict contractual controls.
Assess the task as well as the category. Public information does not make every use low risk, and an approved tool does not make every input acceptable. For client or restricted information, check the applicable controls and permitted use before relying on redaction or an exception.
Map your AI data risks before they become client issues
What to record in an AI input-data register
An input-data register should connect the team, task and tool with the information used and the applicable controls. Use the eight fields below to make those decisions reviewable, not to approve a use merely because its row is complete.
For each AI use case, record:
- the business task, such as drafting, research, analysis, triage, coding, summarisation or client service support;
- the tool or vendor, including whether it is public, enterprise-managed or embedded in another platform;
- the input data categories used;
- whether personal data, client confidential data, financial data or privileged material is involved;
- the retention and training settings you rely on;
- the accountable owner;
- the required human review step;
- the evidence kept for audit or management review.
This does not need to be complex. A spreadsheet, workflow register or governance tool can work if it is maintained and reviewed. The key is that leaders can answer the question without asking every employee to reconstruct their behaviour from memory.
How professional firms should handle client and regulated data
For legal, financial services, insurance and accountancy work, assess the proposed input against the professional, contractual and privacy requirements relevant to the engagement. Assign unresolved questions to the appropriate reviewer.
Route the proposed activity to the relevant legal, privacy, financial or insurance reviewers. Give them the actual input data, task, recipients and decision effects so they can identify applicable requirements and unresolved questions; do not assign duties from a sector label alone.
If the team cannot establish that the proposed input and tool are permitted, pause that use and follow the approval route. Base the decision on the information and restrictions, not on whether disclosure would feel uncomfortable.
Put recurring AI governance checks on a sensible cadence
Data controls that work in day-to-day teams
Give staff instructions they can use at the point of work. The controls below connect tool choice and data handling with review, exceptions and a record of what happened.
- Approved-tool list: name the tools staff may use and the data categories allowed in each one.
- Redaction rules: give examples of what must be removed before prompting, including names, IDs, account numbers and privileged detail.
- Prompt examples: show safe prompts for common tasks, not only banned prompts.
- Access controls: limit sensitive AI workflows to trained roles and enterprise-managed accounts.
- Output review: require human checks for accuracy, bias, confidentiality and client suitability.
- Exception route: provide a named approval path when a team has a legitimate reason to use higher-risk data.
- Audit trail: keep enough evidence to explain what was done, by whom and under which control.
Name a business owner for the proposed use, with IT, legal, compliance and privacy input where relevant. Technical access and business approval are different checks; neither should stand in for the other.
Common mistakes when firms ask staff to use AI
Do not assess the input from the tool category alone. The same assistant may be used for public information, client documents, complaint histories or board papers; assess each task, information type and output use before approving it.
Other mistakes include:
- approving a tool without checking retention, training and admin settings;
- letting teams paste client data into free accounts because the paid tool rollout is slow;
- focusing only on prompt quality while ignoring input-data sensitivity;
- forgetting that embedded AI features in existing software can create new data flows;
- failing to document who reviewed and accepted the output.
Connect the approved-tool list with data categories, review steps and evidence. Record what is allowed and what remains unresolved so staff do not have to infer permission from a tool being available.
Conclusion: answer the data question before the model question
Before choosing a model, buying another AI tool or writing a wider policy, answer the data question. What is going in, who controls it, where does it go, what settings apply and who checks the result?
Record gaps in those answers and assign owners before approving the affected use. Clear records help the team explain the permitted inputs and required controls, but they still need checking against what the workflow actually does.
Turn AI policy into practical workflow controls
Need a safer AI route for your firm?
If this article reflects a live decision in your practice, the useful next step is to map the workflows, confidentiality risks, supplier controls and governance gaps before tools spread informally.
