A law firm's AI decision should begin with the work, the data and the person responsible for the result. A tool may help draft an outline or organize public material, but that does not set permission to upload a client file. Partners should define the data limits and review expectations before staff rely on an unfamiliar workflow. IT can support the resulting decisions through access, policy and IT controls.
The American Bar Association's summary of Formal Opinion 512 addresses competence, privacy, communication and reasonable fees when lawyers use generative AI. Firm leadership should ask qualified advisers how those professional requirements apply to the firm's work. The operating advice here do not set a legal interpretation or certify a firm's compliance.
Define a Workflow Before Approving a Tool
Ask the person requesting AI access to describe the intended task. Find the data needed, the account proposed, the expected output and the attorney responsible for reviewing it. The task gives the discussion a specific purpose. Approval for one use should not quietly become approval to enter every kind of client data into the same app.
Distinguish public data from private matter data. An illustrative public drafting exercise is different from a workflow that uses correspondence, client records or privileged material. The distinction should appear in policy and training examples. Staff should understand what they may use and what requires an extra decision rather than relying on a broad instruction to avoid private data.
Keep a record of the approval. Include the tool and account, users involved, permitted data and review tasks. A partner should be able to explain what was authorized and why. When the work changes, revisit the approval rather than assuming an older decision covers a materially different task.
Set Matter-Specific Data Limits
The firm should find which data may enter a prompt or upload and which must remain outside the workflow. Include client confidentiality, privilege and any matter-specific limits in that decision. Do not make the IT team infer those judgments from a document name. The responsible attorney or data owner should explain the limit before controls are designed around it.
Review attachments as part of the same interaction. A prompt may omit the client's name while an uploaded file contains details that identify people. Removing a visible name does not settle whether the proposed use is permitted. Staff need a process for checking the whole submission and asking for approval when the required data is outside the standard policy.
Document the exception process. Find who may approve a different use, what they need to consider and how the decision is recorded. Any exception should have a clear reason and a clear limit. It should not become an informal route around the firm's data rules. Keep the current instructions accessible so staff can find them before they submit material.
Understand Current AI Use
Leadership may have approved a particular tool without knowing whether staff use other apps or personal accounts. Treat that as a question to investigate rather than a fact to assume. Ask which AI tools are in use, who uses them, how often and what categories of company data enter prompts or uploads. Define the users and devices included in any monitoring engagement.
Tell staff what the review will look at, and why, before it begins. Decide who reviews the findings and how those findings will inform policy and control decisions. The report should connect observations with leadership's questions. It should distinguish current use from a advice about whether that use should continue.
DataSolved's paid 30-day AI Usage Discovery includes monitoring, a report for leadership and a review meeting. The report covers approved and unapproved tools, usage patterns, data categories, risk ranking and recommended next steps. Pricing is confirmed after a short call. It provides evidence for governance decisions without replacing the firm's professional judgment.
Connect Rules with Supported Controls
An approved-tool list should find permitted accounts and tasks. IT controls can support data limits through warnings or blocks for private data in supported AI prompts and uploads, limits on unapproved apps and reporting. Confirm the supported setup and test the intended behavior before describing a control as effective for the firm's workflow.
Plan for legitimate work as well as limits. If staff need an approved way to complete a task, provide a route to request it. A control that blocks intended work should lead to review by the policy owner. A warning should explain the relevant limit and the expected staff member response. Give each issue an owner rather than leaving staff to decide how to bypass the problem.
Regular review keeps the rules connected to new matters, staff roles and tools. The person responsible for governance should maintain the approved list and review exceptions. Partners should know which decisions belong to the firm and which setup tasks belong to IT. Keep legal decisions with firm leadership and its advisers, rather than leaving them to an IT setting.
Prepare Microsoft 365 Before Copilot Use
If the firm is considering Microsoft Copilot, review permissions and oversharing before expanding use. Microsoft states in its privacy records that organizational data is surfaced within existing user permissions. That makes the current access structure relevant to readiness planning. It does not show that every current permission is suitable for the firm's matters.
Ask data owners to review broad access, workspace membership and the intended audience for private material. Plan changes with the responsible attorney or owner involved so approved work can continue. Review sensitivity labels and data hygiene as part of the same readiness discussion. Document which conditions need attention before the proposed workflow begins and who owns the remaining work.
Use training examples that fit the firm's approved data rules. Do not use a private client file merely because it would make a demonstration more realistic. The training task should teach both useful operation and the data limit. Staff should practice finding material that requires another approval before submission.
Keep Review with the Responsible Professional
Assign a qualified reviewer for the output and define what the review should address. The attorney responsible for the work should decide whether the content is accurate, suitable and fit for its intended use. If an important statement cannot be checked, obtain another source or change the workflow before relying on it. The person using AI remains responsible for the final work.
Discuss what data must be retained about an approved workflow with the firm's qualified advisers. The operating process should support the firm's own recordkeeping and client commitments. Avoid assuming that a particular app automatically meets every need. Check the arrangement and document the decision relevant to the proposed task.
See It. Govern It. Use It Well. Set visibility, define limits and support approved work with human review. Schedule an AI Discovery Scoping Call to discuss the paid AI Usage Discovery. Explore AI governance for DFW businesses alongside our law firm IT services.
Connect This Guidance to Your Business
See how these steps fit your business. Start with AI governance and security and AI governance and adoption. Our guide for DFW law firms covers access, staff tasks and client work.
Put Your AI Policy in Writing
Use the AI acceptable-use policy template for small businesses to document approved use, data limits and staff tasks. Adapt the draft to your business and review it with counsel before adoption.
Free Checklist for Business Owners
AI Acceptable Use Policy TemplateA starting policy for how your team uses ChatGPT, Claude, Copilot and Gemini.
