# Template Firm AI Policy > Use when a law firm needs a ready-to-adopt AI use policy governing how associates, paralegals, and staff may use Louis or other AI tools. Covers permitted uses, prohibited uses, mandatory safeguards (citation verification, matter-file documentation, bar-rule compliance), and links to relevant professional responsibility rules. Adaptable for MENA and common-law jurisdictions. - Skill: `sboghossian-mini-claude-for-legal/template-firm-ai-policy` (Agent Skill) - Install (CLI): `npx skillmds@latest add sboghossian-mini-claude-for-legal/template-firm-ai-policy` - Raw SKILL.md: https://api.skillmd.com/api/skills/sboghossian-mini-claude-for-legal/template-firm-ai-policy/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Docs & Writing - License: MIT - Author: sboghossian (https://skillmd.com/u/sboghossian-mini-claude-for-legal) - Updated: 2026-09-10 - Page: https://skillmd.com/skills/sboghossian-mini-claude-for-legal/template-firm-ai-policy --- # Template — Firm AI Use Policy ## When to use this Use this template when: - A law firm is adopting Louis and needs an internal policy before rolling out access to associates and paralegals. - An in-house legal team is formalising its AI-governance framework. - A managing partner or GC asks "what policy should we have in place before our team starts using AI?" - Bar-association guidance requires a written AI policy as a condition of compliant AI use. Adapt the template to the firm's jurisdiction, practice areas, and seniority structure. Have a supervising partner or GC review before circulation. ## Policy template --- **[FIRM NAME] — Policy on Use of AI-Assisted Legal Tools** *Version: [date] | Approved by: [Managing Partner / GC]* ### 1. Purpose This policy governs the use of artificial intelligence tools — including Louis (HAQQ Legal AI) and any other AI-assisted legal research, drafting, or review tool — by all lawyers, trainees, paralegals, and support staff at [Firm Name]. It is intended to ensure that AI use is consistent with our professional obligations, client confidentiality duties, and the standards set by applicable bar regulations. ### 2. Permitted uses The following uses are permitted, subject to the safeguards in Section 4: - **Legal research support:** Using AI to identify potentially relevant statutes, regulations, and case law. All citations must be independently verified before reliance. - **First-draft preparation:** Using AI to generate first drafts of contracts, letters, pleadings, or legal memoranda. All AI-generated drafts must be reviewed, edited, and approved by a qualified lawyer before use. - **Document summarisation:** Using AI to summarise lengthy documents (contracts, court bundles, due-diligence materials) as an aid to review. The reviewing lawyer remains responsible for the underlying documents. - **Pre-litigation document organisation:** Using AI to sort, categorise, and extract key data from document sets. Results must be verified by a lawyer or trained paralegal. - **Translation drafts:** Using AI to produce working translations of foreign-language documents. Translation accuracy must be verified by a qualified translator or lawyer with relevant language competency before the translation is relied upon. ### 3. Prohibited uses The following uses are prohibited: - **Submitting AI output as final work product** without attorney review and sign-off. - **Pasting client-confidential information** into any AI tool not approved under the firm's IT and data-governance policy. - **Court filings generated or substantially drafted by AI** without: (a) attorney review; (b) disclosure to the court where required by applicable court rules or professional guidelines. - **Establishing attorney-client relationships via AI** — no AI tool may be used in a manner that creates the impression that a client is receiving independent legal advice from the AI rather than from a named lawyer. - **Sending privileged or confidential content** to any AI tool outside the firm-approved toolset. - **Relying on AI-generated citations without verification** — fabricated citations are a disciplinary and malpractice risk. ### 4. Mandatory safeguards When using AI tools in legal work, all users must: 1. **Verify all citations.** Every statute, regulation, and case law reference generated by AI must be independently confirmed against the primary source before use in any work product or court submission. 2. **Document AI use in the matter file.** A note must be added to the matter file recording: (a) which AI tool was used; (b) for what purpose; (c) the scope of attorney review applied. This supports privilege analysis and professional responsibility documentation. 3. **Comply with applicable bar rules.** In particular: - Duty of competence: a lawyer must have sufficient understanding of AI outputs to exercise independent professional judgment (see [[safety-bar-rule-1-1-competence-ai]]). - Duty of confidentiality: client information may only be processed through approved tools with adequate data-protection terms. - Supervision: supervising lawyers are responsible for AI outputs produced by lawyers under their supervision, to the same extent as they are responsible for other supervised work. 4. **Flag jurisdiction uncertainty.** If an AI tool's jurisdiction coverage for a specific matter is unclear, escalate to a supervising lawyer before relying on AI-generated analysis. ### 5. Approved tools The following AI tools are approved for use under this policy: - **Louis (HAQQ Legal AI)** — approved for drafting, research, review, and summarisation as described above, subject to data-processing terms on file with IT. - [Add other approved tools as applicable] All other AI tools are prohibited for matter work without prior written approval from [IT Security / Managing Partner]. ### 6. Confidentiality and data All AI use must comply with the firm's data-protection and client confidentiality obligations. Do not use personal data of individuals in AI prompts without a lawful basis. For questions on data handling, see [[template-client-data-explainer]]. ### 7. Review This policy will be reviewed annually or when significant changes to AI tools or bar guidance occur. --- ## Jurisdictional adaptation notes | Jurisdiction | Key adaptation | |---|---| | UAE / DIFC / ADGM | UAE Federal Decree-Law on Personal Data Protection applies; ensure approved tools have UAE-compatible DPA. DIFC and ADGM have their own data-protection frameworks (DIFC DP Law 2020; ADGM GDPR-equivalent). | | KSA | Personal Data Protection Law (PDPL) requires consent for personal data processing; AI prompts containing employee/client personal data must comply. Court filing disclosure requirements are evolving — check latest Saudi Bar Association guidance. | | Lebanon | Bar of Beirut has issued informal guidance on AI; no formal rule yet. Apply the general competence-verification standard. | | EU / UK | GDPR applies to personal data in prompts; AI Act may apply to high-risk legal AI uses. Disclose AI use in court where required under Civil Procedure Rules (UK) or equivalent. | ## Related skills - [[template-client-data-explainer]] - [[template-vendor-security-questionnaire-responses]] - [[safety-bar-rule-1-1-competence-ai]]