Most companies need three months to adopt a new AI tool. The first two are usually spent in a kickoff meeting about the kickoff meeting.
You do not need a consultant or a leadership deck. You need a week and a teammate who is drowning. Ruben Hassid wrote the original version of this playbook — he is right about the substance, even if you can hear the lead magnet humming in the background. I am cutting it down to the bones. The full article with all the prompts is on his Substack.
The team plan
Go to claude.com/pricing/team. Minimum five seats, maximum 150. Premium if you use it daily, Standard if you do not. Claude does not train on your data, yes, people still ask. If you are an enterprise, there is a separate qualification page.
Monday: build Projects, not chats
For team use, Projects beat plain Chat. The point of a Project is not to be a place where you talk to Claude. It is a place where Claude already knows you, your work, and your standards before you say a word.
Your team produces roughly the same five things every week. Proposals. Client updates. Meeting recaps. Briefs. Reports. Each one gets its own Project. Each Project gets one gold-standard example, the relevant context (client list, pricing sheet, brand rules), and an instruction block that defines tone, format, and what good looks like.
Spend forty-five minutes. Build three to five Projects. Make them visible to the team but do not invite anyone yet.
Tuesday: write the shortest possible template
When a teammate opens a Project on Friday, they will look at the blank chat and think: what do I type?
You put the answer in front of them. One sentence. One [INPUT] field for their notes or rough draft. Everything else, tone, format, guardrails, already lives in the Project instructions. The template is not where you teach the AI. It is where you remove the friction.
Five minutes per Project. Save it inside the Project itself.
Wednesday: build the receipts
Pick one real task you did manually this week. Run it through your new system. Save the manual version next to the Claude version. That is your before/after.
Adoption is not a feature problem. It is a comparison problem. People do not switch tools because something is described as powerful. They switch when they see a colleague produce something better in a fraction of the time, using their own work as input.
Make the comparison visible. Screenshots are enough. A screen recording is better.
Thursday: convert one person
Not the early adopter. They will figure it out alone and never thank you. Not the loudest sceptic either — too much friction for day one. Pick the colleague who is behind on emails, always in meetings, staying late. The one who actually does not have time.
Sit with them for fifteen minutes. Open the Project. Use their work, not yours. Run the template. Watch their face.
Then add them to the shared Project and ask what other tasks you should template. They are now a co-owner, not a user. That changes how they talk about it in the next meeting.
Friday: roll out, then shut up
Send the team a short message. Lead with one concrete result from Wednesday. Explain the shared Project in one sentence. List three things they can do today. End with a specific suggestion: “Try the meeting-recap template on tomorrow’s standup. Two minutes.”
Then, separately, DM two or three people directly. Your Thursday convert does the same. One message in a channel is an announcement. Three people in DMs is momentum.
At the end of the day, collect questions. Address the most common one on Monday morning. You are not done. You are deployed.
What you actually built
A workspace where Claude already knows your team’s work before anyone types a thing. That is the difference between Claude as a tool and Claude as a teammate. The tool produces generic output. The teammate produces something that looks like your team actually wrote it.
You did this in five days, at roughly forty-five minutes a day, without an IT ticket and without a consultant. Your competitors are still waiting for the kickoff meeting about the kickoff meeting.
