--- title: How to write an AI usage policy description: Write an AI usage policy with this 7-step guide and free template. 63% of breached organizations lacked an AI governance policy, per IBM 2025. canonical_url: https://www.acronis.com/en/blog/posts/ai-usage-policy-template/ locale: en last_updated: 2026-10-27 --- [Acronis Blog](https://www.acronis.com/en/blog/) # How to write an AI usage policy Write an AI usage policy with this 7-step guide and free template. 63% of breached organizations lacked an AI governance policy, per IBM 2025. [![Image 2: Acronis Editorial Team](https://staticfiles.acronis.com/images/content/d85707ae91461544212d39836dc32b42.png)](https://www.acronis.com/en/blog/authors/acronis/ "Acronis Editorial Team") [Acronis Editorial Team](https://www.acronis.com/en/blog/authors/acronis/) 9 min read Last updated 18 August 2026 To write an AI usage policy, define its scope, list approved and prohibited tools, name the data types banned from prompts, write clear rules, align to recognized frameworks and set enforcement. A written policy is the decision layer; pairing it with technical controls is what turns those decisions into protection. ## Step 1: Define the scope and purpose A policy with an unclear scope is difficult to enforce, so define the boundary first. State which tools the policy covers, which employees it applies to, which data types it governs and which use cases are in scope. Decide between a blanket ban and a risk-based policy. A blanket ban is simple but tends to drive use underground, because employees who need AI often find an unmonitored path. A risk-based policy permits sanctioned tools for defined use cases and is usually more enforceable because it works with how people actually operate. Make the scope explicit. If employees cannot tell whether a given tool or task falls under the policy, enforcement tends to collapse into argument. A one-paragraph purpose statement and a clear scope statement remove that ambiguity before the detailed rules begin. If you have already run a shadow AI audit, use it here: the tools and use cases it surfaced are the real-world scope your policy needs to cover. Our guide to [detecting and governing shadow AI](https://www.acronis.com/en/blog/posts/what-is-shadow-ai-and-why-genai-usage-monitoring-matters-for-msps-and-smds/) walks through that discovery step. ## Step 2: Inventory approved and prohibited AI tools Build a two-column list: sanctioned tools on one side, prohibited tools on the other. This is the most consulted part of the policy, so make it concrete with named tools rather than categories. Set criteria to approve a tool. Review the vendor's data handling terms, the access controls available such as single sign-on and role-based access, and whether the vendor trains models on user data by default. A tool that trains on your prompts without an opt-out generally belongs in the prohibited column until that changes. Handle partially sanctioned tools explicitly. A tool may be approved on a paid enterprise plan but prohibited on its free tier, or approved for one department and not another. Spell out the condition so the distinction is enforceable, and note who to ask when a tool is not on either list. ## Step 3: Define which data types are prohibited in prompts Name the data that may never enter a prompt or upload. Most policies restrict four categories: personally identifiable information, protected health information, payment card data covered by PCI-DSS, and confidential business content such as contracts, financials and strategy. Define each category clearly enough that an employee can apply the rule without asking IT. Give a one-line definition and a concrete example for each, so the line is obvious in the moment an employee is tempted to paste. Call out source code and credentials with their own explicit prohibition. Employees often do not classify code or API keys as sensitive, yet pasting either into a public AI tool can expose intellectual property or hand over live access. Name them directly rather than hoping they fall under a general category. ## Step 4: Write the rules clearly and concisely A rule employees cannot understand is a rule they are less likely to follow. Write in plain language, in short sentences, using the data labels you defined earlier. Distinguish rules from guidelines. A rule is mandatory and carries a consequence; a guideline is advice. Mixing them makes employees unsure what is enforceable, so mark each statement clearly. Avoid vague language, undefined terms and long legal paragraphs no one reads. Here is a poorly written rule: "Employees should exercise appropriate caution and good judgment when leveraging artificial intelligence solutions in the course of their duties." It defines nothing and enforces nothing. The rewrite: "Do not paste client records, source code, passwords or API keys into any AI tool that is not on the approved list." It names the data, names the tools and states the action. ## Step 5: Align the policy to recognized frameworks Framework alignment strengthens a policy operationally and helps demonstrate diligence. It signals to regulators, clients and auditors that your controls follow recognized practice, and it gives you a tested structure instead of a blank page. Alignment is not the same as legal compliance: these frameworks are widely referenced but, on their own, do not satisfy any specific law or regulation, and the applicable requirements depend on your jurisdiction and industry. Anchor governance to the NIST AI Risk Management Framework (AI RMF 1.0), released in January 2023. It is voluntary and non-certifiable, but its GOVERN function --- which is cross-cutting across the framework --- focuses on integrating trustworthy-AI practices into organizational policies, processes and oversight, which makes it a natural reference for the governance section of your policy. Because it is not a certification, you self-attest alignment rather than being audited against it. For a certifiable counterpart, reference ISO/IEC 42001:2023, the first international AI management system standard, published in December 2023. Its leadership requirements (Clause 5) call on top management to establish a documented AI policy, which is why many organizations use NIST AI RMF for the risk vocabulary and ISO/IEC 42001 for an auditable management system. Referencing both by their full names on first mention shows your policy sits inside an accountable structure rather than standing alone. ## Step 6: Set consequences and an enforcement mechanism A policy without consequences is effectively a guideline. State what happens when a rule is broken, or the rule has little force. Make consequences proportionate. A first minor breach might warrant a warning and retraining, while repeated or serious breaches, such as deliberately exfiltrating client data through an AI tool, escalate to formal disciplinary action. Match the response to the severity so the policy is seen as fair and is applied consistently. State the enforcement mechanism inside the policy itself. Name who monitors compliance, how breaches are detected and reviewed, and who decides on consequences. Enforcement that lives only in someone's head is hard to sustain, and employees are more likely to comply with rules they know are actually monitored. This is where a written rule needs a technical control behind it --- monitoring prompt content or AI tool access at the endpoint is what makes a prohibited-data rule observable rather than aspirational. ## Step 7: Review and update the policy regularly AI tools change faster than annual policy cycles, so a once-a-year review can leave the policy out of date for much of the year. Build in a faster rhythm. A practical cadence is quarterly for the approved and prohibited tool list, which changes most often, and annual for the full policy. The tool list is where new entries and new risks tend to appear first, so it needs the more frequent touch. Define what triggers an out-of-cycle review: a major new AI tool entering wide use, a regulatory change such as new AI rules in your jurisdiction, or a security incident involving an AI tool. Naming these triggers means the policy updates when reality changes, not only when the calendar says so. ## AI usage policy template _Copy the template below and replace the bracketed placeholders. It is unbranded and works as a neutral starting point for any organization. It is a drafting aid, not legal advice --- have legal or compliance review it before you publish._ AI Usage Policy Policy name and version: [Organization name] AI Usage Policy, v[1.0], [date] Scope (who it applies to): This policy applies to [all employees, contractors and third parties] using AI tools for [organization] business, on [corporate and personal] devices, covering [standalone AI tools and AI features inside approved SaaS apps]. Purpose (one sentence): To enable safe and productive use of AI tools while protecting [organization] data, clients and compliance obligations. Approved AI tools (list with data handling note): * [Tool name] - [plan/tier] - [data handling note, e.g., does not train on our data; SSO required] * [Tool name] - [plan/tier] - [data handling note] * To request a tool not listed here, contact [owner/role]. Prohibited uses (data types not permitted in prompts or uploads): * [Personally identifiable information (PII)] * [Protected health information (PHI)] * [Payment card data (PCI-DSS)] * [Confidential business content: contracts, financials, strategy] * [Source code and credentials/API keys] * [Any AI tool not on the approved list] Employee responsibilities: * Use only approved tools for [organization] work. * Never enter prohibited data types into any AI tool. * Report suspected data exposure to [contact] within [timeframe]. Consequences for non-compliance: * First minor breach: [warning and retraining]. * Repeated or serious breach: [formal disciplinary action up to and including termination]. Review date and owner: * Tool list reviewed [quarterly]; full policy reviewed [annually]. * Out-of-cycle review triggered by [major new tool, regulatory change or incident]. * Policy owner: [name/role]. Next review date: [date]. ## How Acronis GenAI Protection helps enforce your AI usage policy A written policy tells employees what not to do, but without a technical control at the endpoint it is harder to prevent them from doing it. The gap between policy and enforcement is where data can leak. [Acronis GenAI Protection](https://www.acronis.com/en/products/cloud/cyber-protect/genai-protection/) helps close that gap by turning parts of a written policy into working controls where AI use happens. It discovers and monitors browser-based generative AI tool use, giving you the shadow AI visibility that shows whether the policy is being followed. It can inspect prompts for personally identifiable information, protected health information, PCI-DSS data and content marked confidential, in detect-only or block mode, turning your prohibited-data rules into enforced controls. It also detects and blocks harmful prompts, including prompt injection. Where a tool needs to be stopped, unwanted AI usage can be blocked through URL filtering via the Acronis EDR or XDR integration. Together these map the discovery, data and enforcement sections of a policy onto working controls. GenAI Protection focuses on browser-based AI use, so it is strongest when paired with complementary controls. [Acronis Data Loss Prevention](https://www.acronis.com/en/products/cloud/cyber-protect/data-loss-prevention/) extends coverage to sensitive data leaving through 70-plus local and network channels, including local GenAI apps, and [Acronis Security Awareness Training](https://www.acronis.com/en/products/cloud/cyber-protect/security-awareness-training/) helps employees internalize the policy so fewer violations happen in the first place. Acronis GenAI Protection runs on Windows 10, Windows 11 and macOS, delivered through the Acronis platform, so policy enforcement sits alongside existing endpoint protection rather than in a separate console. ## Put your AI usage policy into force A policy is only as strong as the controls behind it. See how Acronis GenAI Protection discovers shadow AI, enforces your prohibited-data rules and blocks harmful prompts across client environments: [book a 1:1 demo](https://www.acronis.com/en/products/cloud/demo/?service=genai-protection) or [start a free trial](https://www.acronis.com/en/products/cloud/trial/). To see where GenAI protection fits in Acronis's wider AI strategy, explore the [Acronis AI-native platform](https://www.acronis.com/en/acronis-ai/). ## Key takeaways * Policy comes first: every technical control enforces a written decision, so the policy is the foundation. * The governance gap is wide: 63% of breached organizations either had no AI governance policy or were still developing one, according to the IBM Cost of a Data Breach Report 2025. * Aligning to the NIST AI Risk Management Framework 1.0 and ISO/IEC 42001:2023 strengthens the policy operationally and signals diligence, but neither framework is itself a legal requirement. * The template is a starting point to copy and adapt, not a finished or legally reviewed policy for any specific organization. * Enforcement differs from policy: a rule without a consequence and a monitoring mechanism is only a guideline. * Review the tool list quarterly and the full policy annually, with out-of-cycle reviews for major tools, regulatory change or incidents. * A policy is enforceable only when it is paired with technical controls that can observe and act on AI use at the endpoint. ## Frequently asked questions **How long should an AI usage policy be?** Short enough that employees read it, usually two to four pages. The approved tool list, the prohibited data types and the consequences are the parts people consult, so keep those clear and put detailed background in an appendix. A concise policy that is read and followed beats a comprehensive one that is ignored. **Do I need a lawyer to write an AI usage policy?** You can draft it without one, but have legal or compliance review it before publishing, especially the consequences, data handling and regulatory sections. A lawyer helps ensure the enforcement provisions are valid in your jurisdiction and that the policy aligns with employment law and data protection obligations. The draft is yours; the review protects you. This article is general guidance, not legal advice. **What is the difference between an AI usage policy and an AI governance framework?** An AI usage policy is the rules employees follow day to day: which tools are allowed and what data is banned from prompts. An AI governance framework is the wider system of roles, risk processes and oversight that the policy sits inside. The policy is one output of the framework, not a replacement for it. **How do I get employees to actually follow the AI usage policy?** Make compliance the easy path. Provide sanctioned tools that do the job, so employees do not need workarounds. Write rules in plain language they can apply without asking IT. Train briefly and repeat after changes. Then enforce consistently with proportionate consequences, because a policy known to be monitored is more likely to be followed. **What frameworks should an AI usage policy reference?** Reference the NIST AI Risk Management Framework 1.0, a voluntary framework whose GOVERN function anchors AI policy and oversight, and ISO/IEC 42001:2023, the certifiable AI management system standard whose leadership clause calls for a documented AI policy. Together they give your policy a recognized governance structure and an auditable management system that regulators, clients and auditors understand --- though alignment with them is not a substitute for meeting the specific laws that apply to you. [Ransomware protection](https://www.acronis.com/en/blog/tags/ransomware-protection/) [Cyber protection](https://www.acronis.com/en/blog/tags/cyber-protection/) [MSP cybersecurity](https://www.acronis.com/en/blog/tags/msp-cybersecurity/) [Antivirus protection](https://www.acronis.com/en/blog/tags/antivirus-protection/) [Corporate cybersecurity](https://www.acronis.com/en/blog/tags/corporate-cybersecurity/) [Cyber protection for businesses](https://www.acronis.com/en/blog/tags/cyber-protection-for-business/) [Security software for business](https://www.acronis.com/en/blog/tags/security-software-for-business/) ## Written by ![Image 3: Acronis Editorial Team](https://staticfiles.acronis.com/images/content/d85707ae91461544212d39836dc32b42.png) Acronis Editorial Team Acronis is a global cyber protection company that provides natively integrated cybersecurity, data protection, and endpoint management for managed service providers (MSPs), small and medium businesses (SMBs), and enterprise IT departments. [Previous post](https://www.acronis.com/en/blog/posts/ot-vs-it-comparison/) [OT vs IT: The Key Differences Between Operational Technology and Information Technology](https://www.acronis.com/en/blog/posts/ot-vs-it-comparison/) [Next post](https://www.acronis.com/en/blog/posts/best-email-security-solutions-for-msps-2026/) [Best email security solutions for MSPs in 2026: a buyer's guide](https://www.acronis.com/en/blog/posts/best-email-security-solutions-for-msps-2026/) About Acronis Acronis is a leader in AI-native cyber protection, unifying data protection, cybersecurity, IT operations and hybrid infrastructure to deliver business continuity, security and compliance with less cost and complexity. More than 21,000 service providers trust Acronis to protect over 750,000 businesses in 150 countries. Acronis is headquartered in Schaffhausen, Switzerland. [Add as a preferred source on Google](https://google.com/preferences/source?q=acronis.com) ## Related stories Want to learn more? Read the latest stories featuring this solution on our blog. [Acronis Blog](https://www.acronis.com/en/blog/) ![Image 4: stc Bahrain and Acronis partner to bring advanced endpoint security to businesses in Bahrain](https://staticfiles.acronis.com/images/blog-card/9600f3eea83b8f4e168ea6c1409c8c26.png) [stc Bahrain and Acronis partner to bring advanced endpoint security to businesses in Bahrain](https://www.acronis.com/en/blog/posts/stc-bahrain-and-acronis-partner-to-bring-endpoint-security-to-businesses/) We are proud to announce a strategic partnership with stc Bahrain, expanding access to advanced cybersecurity services for businesses across Bahrain. **October 26, 2026**3 min read ![Image 5: Acronis expands MDR toward broader workspace protection](https://staticfiles.acronis.com/images/blog-card/18e90e89f5ec4ef0901f19008d4ccbf2.png) [Acronis expands MDR toward broader workspace protection](https://www.acronis.com/en/blog/posts/acronis-expands-mdr-toward-broader-workspace-protection/) Acronis is expanding the MDR service with capabilities that make this expertise easier to deliver and help partners address more risks. The developments build on a broader commitment to managed security: helping service providers protect the connected workspace while giving customers access to the expertise needed to investigate threats, respond effectively and maintain business continuity. **October 13, 2026**6 min read ![Image 6: Acronis Cyber Protect v17: Improving cyber resilience across people, data and infrastructure](https://staticfiles.acronis.com/images/blog-card/c2467c28c0278609eb3560cb2a8a4d53.png) [Acronis Cyber Protect v17: Improving cyber resilience across people, data and infrastructure](https://www.acronis.com/en/blog/posts/acronis-cyber-protect-v17-improving-cyber-resilience-across-people-data-and-infrastructure/) Acronis Cyber Protect v17 Q3 2026 expands integrated cyber protection with new capabilities that strengthen security awareness, improve backup operations and extend protection across more virtual environments. The result is stronger cyber resilience, simplified operations and broader protection delivered through a single platform. **October 05, 2026**4 min read