# How should an IT team roll out ChatGPT or Copilot to the whole company? > Pick one assistant, lock down the tenant and data settings, fix oversharing, pilot with a small group, then roll out with a policy. Free lessons. Source: https://learn.ynteractive.com/content/questions/how-it-teams-roll-out-ai-safely · Updated 2026-10-21 · Free from Stepthrough (https://learn.ynteractive.com) By [Robert Breen](https://learn.ynteractive.com/content/about) · Published October 21, 2026 · Updated October 21, 2026 · 4 min read Roll it out in five steps: choose one approved assistant, set up the business tenant with [single sign-on](https://learn.ynteractive.com/content/glossary/single-sign-on) and the right data settings, clean up file oversharing before Copilot can surface it, run a pilot with a small group, and then launch company-wide with a one-page [AI policy](https://learn.ynteractive.com/content/glossary/ai-policy) and short training. The settings people most often ask about, and the team policy that goes with them, are covered in [ChatGPT: Keep Client Data Safe](https://learn.ynteractive.com/content/chatgpt-data-controls-protect-client-data). The reason to move deliberately is simple. Your staff are probably already using AI on personal accounts. A good rollout gives them a better, safer option quickly, rather than a ban they will work around. ## Step 1: Which assistant should we pick? Usually the one that fits your existing stack: - **Microsoft 365 shops**: [Microsoft 365 Copilot](https://learn.ynteractive.com/content/glossary/microsoft-copilot), which works inside Word, Excel, Outlook, PowerPoint and Teams. - **Google Workspace shops**: Gemini for Google Workspace, built into Gmail, Docs, Sheets and Meet. - **Either**: a ChatGPT or Claude business plan, when teams need strong general writing and analysis. Test it on a real task first. In [Choose the AI Assistant Your Team Should Use](https://learn.ynteractive.com/content/chatgpt-vs-claude-vs-gemini-vs-copilot-same-task), the same client reminder email and task table go to ChatGPT, Claude, Gemini and Copilot. All four pass on the facts, and the office manager picks Copilot Chat because the agency already pays for Microsoft 365. ## Step 2: How do we set up the tenant? - **Business plan only.** OpenAI says it does not train on ChatGPT Business or Enterprise data by default. Claude, Copilot and Gemini business plans have their own commercial data terms. Read the current terms for the plan you buy. - **Single sign-on.** OpenAI lists SAML SSO on ChatGPT Business, and Enterprise adds features such as SCIM provisioning and role-based access. Check each vendor's current plan page. - **Admin roles.** Decide who are owners and admins, and who can add connectors or apps. - **Data features.** Decide on memory, shared links, file uploads and third-party connectors before launch, not after. - **Retention and audit.** Know what logs you can export and how long chats are kept, and tell legal. ## Step 3: Why fix oversharing before Copilot? Microsoft's documentation says Microsoft 365 Copilot only surfaces organizational data the user already has at least view permission for. That is good news and a warning. If a salary spreadsheet sits in a folder shared with "Everyone", Copilot will find it for anyone who asks. Microsoft points admins to tools in Microsoft Purview and SharePoint Advanced Management to find overshared, ownerless and sensitive sites, and to options like Restricted SharePoint Search while you clean up. The same logic applies to Gemini and Google Drive sharing. ## Step 4: How do we run the pilot? Pick 15 to 30 people from different departments, including a skeptic or two. Give them two or three real tasks each, a channel for questions, and a short form for what worked and what went wrong. Watch for: - Tasks that save time and tasks that do not. - Data people tried to paste that should not go in. - Settings that confused people. ## Step 5: How do we launch company-wide? 1. **Publish the AI policy.** Approved tools, what never goes in, a person checks every output, and how to report a mistake without blame. [Roll Out a Team AI Policy in One Afternoon](https://learn.ynteractive.com/content/ai-acceptable-use-policy-claude) drafts one from your rules, with an announcement email and a five-question quiz. 2. **Train by role**, not by feature. Work with L&D; see [how to train staff to use AI](https://learn.ynteractive.com/content/questions/how-to-train-staff-to-use-ai). 3. **Block or discourage unapproved tools** once the approved one is available. 4. **Review quarterly.** Vendors change settings and features often. ## What does this look like in a real lesson? In [ChatGPT: Keep Client Data Safe](https://learn.ynteractive.com/content/chatgpt-data-controls-protect-client-data) you are Priya, office manager at a six-person firm, after a client asks in writing what happens to their data and someone has pasted a payroll export. In Settings you turn off **Improve the model for everyone** under Data controls, open **Manage** next to Shared links, see what **Export data** does, turn off **Enable memory** under Personalization, and try **Temporary Chat**. You finish by copying an eight-rule team AI policy. On a business workspace, admins control several of these settings centrally, which is exactly why IT should own the setup. ## What mistakes should IT teams avoid? - **Launching Copilot before an access review.** - **Banning AI with no approved option.** Shadow AI grows. - **Turning on every connector at once.** - **No owner for the policy.** Pair IT with legal, compliance and L&D. - **Promising privacy you have not checked.** Staff ask [whether their employer can see their chats](https://learn.ynteractive.com/content/questions/can-my-employer-see-my-chatgpt). Know the answer for your plan. ## Key takeaways - Choose one assistant that fits your stack, and test it on real tasks. - Business plan, SSO and agreed data settings before anyone logs in. - Fix oversharing before Copilot or Gemini can surface it. - Pilot, then launch with a policy and role-based training. ## Frequently asked questions **Should we allow more than one AI assistant?** Start with one to keep settings, training and support simple. Add a second only for a clear need that the first cannot meet, under the same policy. **How long does a company-wide rollout take?** It depends on size and how much file cleanup Copilot needs. The pilot should run long enough for people to use it on real work, typically a few weeks. **Where can I find IT-specific prompts?** The [prompts for IT managers](https://learn.ynteractive.com/content/prompts/roles/it-managers) cover an AI acceptable use policy, evaluating an AI tool before approving it, help desk replies, outage notices and more. ## Learn it hands-on Free, interactive lessons: you do every step yourself in a practice copy of the real tool, it checks your work as you go, and nothing touches your real accounts. - [ChatGPT: Keep Client Data Safe, Set Up Your Team's Data Controls in 10 Minutes](https://learn.ynteractive.com/content/chatgpt-data-controls-protect-client-data): A client asks if their files go into ChatGPT. Answer with confidence: set the data controls, memory, Temporary Chat and shared links, then hand your… (about 10 minutes) - [Roll Out a Team AI Policy in One Afternoon using Claude](https://learn.ynteractive.com/content/ai-acceptable-use-policy-claude): Your team already uses AI, with or without rules. Have Claude write a one-page acceptable use policy from your rules, check the parts that matter… (about 10 minutes) - [Choose the AI Assistant Your Team Should Use, Tested on ChatGPT, Claude, Gemini and Copilot](https://learn.ynteractive.com/content/chatgpt-vs-claude-vs-gemini-vs-copilot-same-task): Give one real office task to ChatGPT, Claude, Gemini and Microsoft Copilot, check each answer on the facts, and pick the one to use at work. (about 8 minutes) [Start the free lesson](https://learn.ynteractive.com/modules/chatgpt-data-controls) ## AI tools in this answer Each one in the Stepthrough AI tools directory: what it does, who it is for and the free lesson where there is one. [ChatGPT](https://learn.ynteractive.com/content/ai-tools/tool/chatgpt) · [Microsoft 365 Copilot](https://learn.ynteractive.com/content/ai-tools/tool/microsoft-365-copilot) · [Claude](https://learn.ynteractive.com/content/ai-tools/tool/claude) · [Gemini for Google Workspace](https://learn.ynteractive.com/content/ai-tools/llms-chat-assistants#tool-gemini-for-google-workspace) [The full AI tools directory](https://learn.ynteractive.com/content/ai-tools) ## Related terms Plain-English definitions, each with an example from a free lesson. [Single sign-on (SSO)](https://learn.ynteractive.com/content/glossary/single-sign-on) · [AI policy](https://learn.ynteractive.com/content/glossary/ai-policy) · [Microsoft Copilot](https://learn.ynteractive.com/content/glossary/microsoft-copilot) · [AI assistant](https://learn.ynteractive.com/content/glossary/ai-assistant) · [AI incident response](https://learn.ynteractive.com/content/glossary/ai-incident-response) · [AI scorecard](https://learn.ynteractive.com/content/glossary/ai-test-scorecard) [The full AI glossary](https://learn.ynteractive.com/content/glossary) ## Related prompts - [AI assistant answer scorecard prompt](https://learn.ynteractive.com/content/prompts/ai-assistant-answer-scorecard-prompt): A prompt that checks four AI assistants' answers to the same task against your own facts and fills a scorecard, so you can pick… - [ChatGPT memory sort prompt](https://learn.ynteractive.com/content/prompts/chatgpt-memory-sort-prompt): Paste your ChatGPT saved memories and sort each one: keep it, move it to custom instructions, or delete it because it holds… - [IT incident report prompt](https://learn.ynteractive.com/content/prompts/it-incident-report-prompt): A ChatGPT prompt for an IT incident report: turn your log into a timeline, what was affected, how you contained it and what is… [All free prompt templates](https://learn.ynteractive.com/content/prompts) ## Related questions - [How should a finance team start using AI without breaking controls?](https://learn.ynteractive.com/content/questions/how-to-start-using-ai-in-finance): Treat AI as a preparer, never an approver: written rules, an approved tool, a person who reviews and evidence you keep. A plan… - [How should L&D roll out AI training to the whole company?](https://learn.ynteractive.com/content/questions/how-to-train-staff-to-use-ai): Start with a one-page AI policy, then hands-on practice on real tasks, then role-based tracks. An L&D rollout plan with free… - [Can lawyers use ChatGPT without breaking their ethics rules?](https://learn.ynteractive.com/content/questions/can-lawyers-use-chatgpt-ethically): Yes, if you understand the tool, protect client confidences, supervise staff, verify every citation and bill fairly. Not legal… - [How do I convince my boss to let our team use AI?](https://learn.ynteractive.com/content/questions/convince-boss-to-allow-ai): Bring one low-risk use case, a short trial, data rules written down and a demo you ran yourself. A step-by-step pitch, with free… - [How are compliance teams using AI?](https://learn.ynteractive.com/content/questions/how-compliance-teams-use-ai): Compliance teams use AI to compare rule and policy versions, answer questions from their own documents with citations, and write… - [How are school administrators using AI?](https://learn.ynteractive.com/content/questions/how-school-administrators-use-ai): School administrators use AI for family newsletters, board meeting minutes, schedules and policy drafts, keeping student records… [More questions: AI by department](https://learn.ynteractive.com/content/questions/by-department) · [All AI questions](https://learn.ynteractive.com/content/questions) · [All free tutorials](https://learn.ynteractive.com/content)