Claude usually arrives in a company through a few people who use it on their own because it writes well and copes with long documents. Then the company buys Claude Team or Claude Enterprise, and for everyone else it becomes one more chat window.
The seats renew, but the projects stay empty, nobody has built a Skill, and the connectors to file storage and CRM were never switched on.
Claude training for teams, not a course for one person
We sell in-company Claude training for teams. To learn Claude on your own, Anthropic's free Claude Academy courses, such as Claude 101, are the better start. How the options compare (Specialty Tokens is one of the providers and wrote this page):
| Option | Strength | Fits when | Doesn't fit when |
|---|---|---|---|
| Claude Academy (Anthropic, free, online) | Official, current, self-paced | One person wants to learn the features | A team needs shared projects, Skills and its own systems connected |
| Course marketplaces (Coursera, Udemy and similar) | Structured syllabus, many angles | You want a specific topic, such as Claude Code | The recording lags behind Claude's frequent updates |
| Open workshops and live classes | Instructor-led, mixed groups | Two or three people from different teams need a start | The work involves confidential documents or your own systems |
| Large enablement firms | Multi-week programs across roles and levels | A central L&D program for thousands of employees | You want a few departments working in Claude within weeks |
| In-company training (for example Specialty Tokens) | On site, on the team's own work, connectors set up first | A department should change how it works, with real data | Only one or two people need to learn, or you need a certificate |
How our Claude training works
Claude training from Specialty Tokens starts before the first session. We agree which teams take part and which data Claude may see, and set up the connectors to file storage, CRM or ERP. Where no ready-made connector exists, we build the connection through the Model Context Protocol (MCP), the open standard Anthropic created. In the session, participants work on their own tasks, not on examples.
These features turn Claude into a team tool (requirements according to Anthropic, as of October 2026):
| Feature | What it does for a team | Requirement |
|---|---|---|
| Projects | Group files, instructions and chats for a client, product or bid | All plans (Free: up to 5); sharing with colleagues on Team and Enterprise |
| Artifacts | Documents, dashboards and small tools built next to the chat and shared | All plans |
| Skills | Capture recurring workflows, such as a report format or brand guidelines | Code execution enabled; Owners can provision them org-wide on Team and Enterprise |
| Connectors (MCP) | Claude reads from and acts in Google Workspace, Microsoft 365, Slack or your own systems | An Owner enables them, each user signs in; custom MCP servers must be reachable from Anthropic's cloud |
| Research | Multi-step research across the web and connected sources, with citations | Paid plan (Pro, Max, Team, Enterprise), web search turned on |
Typical content:
- Clear instructions instead of prompt tricks: give context, include examples, check the output.
- Long documents: compare contracts, summarize policies, find what changed between versions.
- Skills for the department's workflows, built in the session and provisioned for everyone.
- Research and Cowork: sourced research and multi-step tasks Claude works through on its own.
- Data rules: what goes in, what stays out, and how to handle sensitive content.
Claude vs ChatGPT, Copilot and Gemini for a company
| Tool | Strong at | Fits when | Fits less well when |
|---|---|---|---|
| Claude (Team, Enterprise) | Long documents, careful writing, Skills, MCP connections | Document work dominates, own systems should be connected | Work happens in Outlook, Teams and Excel and Copilot is licensed |
| ChatGPT (Business, Enterprise) | Broad features; familiar to many employees | One general assistant for many departments | The AI should work inside Word, Outlook or Gmail |
| Microsoft 365 Copilot | Working inside Word, Excel, PowerPoint, Outlook and Teams | Microsoft 365 is the workplace | Key data and workflows live outside Microsoft 365 |
| Gemini (Google Workspace) | Working inside Gmail, Docs, Sheets, Meet and Drive | The company runs on Google Workspace | The company runs on Microsoft 365 |
Many companies combine two tools, for example Copilot for everyday office work and Claude for document-heavy work and engineering. The longer version is our ChatGPT vs Copilot vs Claude vs Gemini comparison. We train in the tool you have approved, also as ChatGPT training, Microsoft Copilot training or Gemini training.
What teams build
We train teams together, on site in the US, the UK and across Europe, with separate formats for leadership and teams. Every session ends with a shortlist of workflows worth automating, and the first ones running.
At TimeTac, the marketing and growth team built its own automations after an AI enablement workshop and an internal AI platform: ten live, 117 runs in three weeks, 54 unattended, no engineers on the team. At Speedinvest, the pan-European VC, enablement ran team by team with Affinity CRM and the fund's ERP connected; the dealflow inbox automation saves an estimated 10 to 20 hours a week.
Claude Code for engineering teams
With Claude Code, Claude works directly in the codebase, writes and tests code, and opens pull requests. That has its own format: Claude Code training for engineering teams, with Codex and Cursor in comparison.
Everything built in the training is yours: the projects, the Skills, the workflows. No per-seat fees to us and no lock-in. This training is part of our AI training for employees.
