You bought Claude Teams six months ago. Three people on the team use it every day. Two more dabble. Everyone else logged in once, tried a prompt, got a generic answer, and quietly went back to what they were doing before. The license count on the invoice says 25. The active-user count in your admin panel says seven.
That gap is the 30% adoption plateau. It is the single most common pattern I see across the teams I work with — agencies, VC firms, product companies, professional services. Different industries, same plateau. And the pattern is almost never about the people. It is about how the rollout was structured.
Here are the five reasons your rollout has stalled — and the specific move that gets it unstuck for each one.
Why this stall is so common (and so misunderstood)
The standard reaction to a stalled rollout is to blame the team. "They are not curious enough." "They are stuck in their old ways." "They need to be more open to AI."
That framing is wrong, and it costs you the rollout. Every team that hit 70%+ adoption did so because leadership stopped treating Claude as a tool announcement and started treating it as a workflow change. The teams stuck at 30% are stuck because the rollout was structured as if Claude were Slack — install it, send an email, expect usage.
Claude is not Slack. Slack replaces email threads. Claude replaces ways of thinking. Those require completely different rollout strategies.
If your rollout is stalled, one of the five patterns below is almost certainly the cause. Possibly two. Read all five before deciding which applies to you, because the symptoms overlap.
Reason 1 — You skipped the foundation week
Most teams skip the first week of a real rollout entirely. They go straight to "everyone log in and try it." The result is twenty-five people exploring a powerful tool with no shared context, no shared examples, no shared goal. Three of them figure it out on their own. The other twenty-two get a bad first impression and never come back.
The foundation week is not exciting. Nobody writes a LinkedIn post about it. But it is the difference between a rollout that compounds and one that flat-lines.
The foundation week looks like this: admin setup (SSO, billing, role assignments, usage policies), leadership alignment (what does success look like in 60 days, who owns this rollout, what is the budget for outside help), and a dedicated Slack or Teams channel for Claude questions. Three concrete artifacts. One week.
If you skipped this, your rollout was always going to plateau. The fix is to back up and do it now, even if you are already three months in. Resetting the foundation costs five days. Not doing it costs the entire rollout.
The full 60-day rollout sequence — including the foundation week and what to do after — is in How to Roll Out Claude to Your Team: A 60-Day Adoption Plan.
Reason 2 — Your team is using generic prompts on generic problems
Generic training gets 23% adoption. Role-specific training gets 67%. That is not my number — it is the industry benchmark, and I see it confirmed in every engagement.
Here is the pattern: someone shows the team a few generic prompts at kickoff. "Summarize this document." "Rewrite this email more professionally." "Brainstorm five ideas for our launch." The team tries them. Claude does what it does. The output is fine. Nobody is impressed, because the output could have come from any AI tool. The team concludes that Claude is just another chatbot and goes back to whatever they were doing.
That conclusion is wrong, but it is the rational conclusion based on the demo they saw. Generic prompts produce generic value, and generic value does not change behavior.
The fix is to map Claude to the actual workflows your team runs every week. Your finance lead does not need to learn how to "brainstorm five ideas." They need a prompt template that turns a board memo draft into a CFO-ready deck. Your operations manager does not need "rewrite this email." They need a Skill that takes a new-hire intake form and generates the first-week onboarding plan in their voice.
Workflow-specific prompts produce workflow-specific value. That is when adoption breaks 50%.
The fastest way to get role-specific prompts in front of your team is to spend a paid month learning your business well enough to build them. That is what the Clarity Strategy Session is. Or you can do it yourself if you have the time — start by interviewing one person on every functional team about their three most repetitive weekly tasks, and build prompts for those.
Reason 3 — The power user tax
Every stalled rollout has one to three power users. Usually the person who championed the purchase, plus one or two early enthusiasts. They love Claude. They have built Projects, configured Skills, set up MCP connectors. They are getting real value.
And they are silently carrying the rollout. The CEO asks them to demo something at the next all-hands. New hires get pointed to them. Anyone with a Claude question routes through them. They become the de facto Claude help desk, on top of their actual job.
This is the power user tax, and it kills rollouts in two ways. First, the power users burn out and stop spreading knowledge because they are exhausted. Second, the rest of the team learns that Claude is something only the power users do. The mental model becomes "Claude is the thing those people use," not "Claude is how we work now."
The fix is to formalize what the power users have already figured out and remove the bottleneck. Document their best Projects. Turn their best prompts into reusable Skills the whole team can access. Run a 90-minute team workshop that walks everyone through the system the power users built, so the team can use it without asking. The power users get their time back. The rest of the team gets a working system instead of a single person to ask.
A Foundations Workshop is the formal version of this — a 90-minute live training session where the team learns Claude basics together with role-relevant examples. A Custom Workshop goes a layer deeper, built around your actual workflows.
Reason 4 — Nobody is watching Claude for you
Claude ships meaningful new features every two weeks. Skills shipped, MCP connectors expanded, Projects gained shared instructions, Cowork launched, scheduled tasks became real, Claude Code matured into something most CEOs are using daily. Six months ago, half of these did not exist.
Most teams I work with set up Claude once and never touched the architecture again. The Projects they built in January are not using the connectors that shipped in March. The Skills they wrote do not take advantage of the long-context improvements from May. Their rollout is technically still running, but it has been frozen in time since the day they last sat down to configure it.
This is the slowest-moving reason rollouts stall, and the hardest to see. The team is still using Claude. The numbers do not collapse. They just stop growing. Adoption flat-lines at whatever level it hit in month two, because the system never gets better.
The fix is to have someone whose job — formally or informally — is to monitor Claude releases and tell the team what matters. This is what most of my Partnership clients pay for. It is also a role you can run yourself: subscribe to the Anthropic changelog, block thirty minutes every other Friday, decide what the team should know about, and send a one-paragraph internal note. The discipline is more important than the duration.
If you cannot commit thirty minutes every two weeks to this, you are not actually running a Claude rollout. You are running a Claude purchase.
Reason 5 — There are no behavior triggers
The most invisible reason rollouts stall is also the most fixable. Your team has not stopped using Claude because they decided it was not useful. They have stopped using Claude because nothing in their daily workflow reminds them to use it.
When a Slack message comes in, the trigger is to reply in Slack. When a calendar invite arrives, the trigger is to accept or decline. When a Linear ticket gets assigned, the trigger is to open Linear. These triggers are everywhere. They make those tools sticky.
Claude has almost no native triggers. It is a tab in the browser. If you do not open the tab, nothing in your day will remind you that Claude exists. The default workflow does not include it.
The fix is to engineer the triggers deliberately. A few that work for almost every team:
- A morning check-in skill. Forward-slash command, runs once at the start of the day, surfaces priorities, calendar, and any decisions waiting on you. Free templates exist on claudeforeveryone.com.
- Pinned Project per active client. Every client gets a dedicated Project with their context loaded. When someone needs to think about that client, the trigger is to open the Project — not to start a new chat.
- A weekly meeting-prep ritual. Thursday afternoon, every team member feeds the next week's calendar into Claude and asks for prep notes. Friday morning, they review. Monday they walk in ready.
- Forward-slash commands for the three most repeated tasks per role. Skills do not always auto-trigger reliably from natural language. Typing /morning or /standup or /checkin works every time.
Triggers are the difference between a tool you have and a tool you use. Engineer them, or the rollout will keep stalling no matter how good the architecture is.
How to figure out which reason applies to you
If you read through those five reasons and felt called out by more than one, that is normal. They usually cluster. The fastest way to figure out which is the dominant problem is to ask three people — your highest power user, one moderate user, and one near-zero user — the same two questions:
- When was the last time you used Claude for something other than what we showed you at kickoff?
- What would have to be true for you to use it three times this week?
The answers map directly to the five reasons. Power users almost always answer Reason 4 (the system is stale). Moderate users answer Reason 2 or 5 (no role-specific prompts, no triggers). Near-zero users answer Reason 1 or 3 (no foundation, power-user dependent).
You do not need a strategist to ask these questions. You need to actually ask them, and act on what you hear.
What "unstuck" looks like
A rollout that has cleared the 30% plateau looks like this: every functional team has at least two role-specific prompts they use weekly. New hires get a fifteen-minute Claude onboarding as part of their first day. Someone is watching for Claude updates and the team knows who. The power users are mentors, not bottlenecks. People talk about "what Claude said about that" the way they used to talk about "what the data showed."
You do not get to that state by buying more seats or sending another all-company email. You get there by treating the rollout as a real piece of operational work — same rigor as launching a new product, hiring a new function, or onboarding a new client.
If you want help, the path is the Clarity Strategy Session. One month, fully embedded, leaves you with a blueprint for what to build, in what order, with projected ROI per recommendation. If you want to keep someone in the room after that to make sure the rollout actually lands, the Partnership starts at $1,800/mo.
If you would rather start with a 90-minute team training session — no embedding, no blueprint, just the team learning Claude together with role-relevant examples — that is the Foundations Workshop.
And if you are the power user reading this and you want to set up the system properly for your own team without an embedded engagement, the Quick Start 1:1 is the right move. Three sessions, $1,247, you walk out with a working system you understand and can keep building on your own.
The 30% plateau is not a tool problem. It is a system problem. Systems can be fixed.
Nicole Patten is the founder of Elevate Online and runs a Claude-specific training practice. She spent 7 years at Google as a Senior UX Engineer before dedicating her career to helping teams use AI responsibly and effectively. 100% of her business runs on Claude.