Giving 50 people access to Claude is easy. Getting 50 people to use it well is not. A structured rollout makes the difference between a tool a few enthusiasts love and a capability the whole company relies on.
Key takeaways
- Set up admin, data settings and connectors before inviting users.
- Start with a pilot group of 8–12 people across departments.
- Build shared Projects and skills for each team’s real work.
- Train by role, appoint champions and track adoption.
Step 1: Decide the basics
Agree on why you’re rolling out Claude, which plan fits, who owns the rollout and what success looks like after 90 days. Typical goals include saving time on writing and research, improving the quality of customer communications and reducing repetitive admin.
Step 2: Configure before you invite
- Set up the organization, admin roles and, where available, single sign-on.
- Review data, retention and sharing settings against your policies.
- Connect the tools people use most, such as email, calendar, documents, CRM and chat, where connectors are available.
- Publish a short acceptable-use policy explaining what data can be used and when human review is required.
Step 3: Run a pilot
Choose a pilot group from marketing, sales, operations, finance and leadership, including both enthusiasts and skeptics. For two to four weeks, have them use Claude on real work, capture what works and note where they get stuck.
Step 4: Build shared Projects and skills
Turn what the pilot learned into shared resources. A marketing Project might hold positioning, brand guidelines and example content. A sales Project might hold product information, objection handling and proposal templates. Skills package repeatable processes, such as writing in your brand voice or preparing a weekly report, so they run consistently.
Step 5: Roll out in waves
Expand department by department. Each wave gets a kickoff, a role-specific workshop, its shared Projects and skills, and a named champion. Waves of 10–15 people keep support manageable.
Step 6: Measure and improve
Track active users, usage by team and the workflows people rely on. Collect short success stories and time-saving estimates. Hold monthly office hours and update Projects and skills as you learn.
Common mistakes to avoid
- Inviting everyone at once with no training.
- Skipping data settings and usage rules.
- Treating AI as an IT project instead of a change in how people work.
- Measuring logins instead of outcomes.