Free preview. This lesson is part of a paid course that earns the ETA Specialist: Claude credential. About this course
Welcome: Claude at work, the ADAPT way
In this lesson you'll
- Describe what Claude can and can't see in a chat, a Project and a connected app
- Check your plan's data and training settings before you share work content
- Apply ADAPT to decide whether a task is a good fit for Claude
Welcome
Welcome to Claude at Work from the Enterprise Technology Association (ETA). Claude is an AI assistant made by Anthropic. People at work use it to draft and edit, summarize long documents, analyze data, write code, research and think through problems. This course is about using it well: getting work you can trust, keeping data safe, and showing real value.
This course assumes you've completed Level 1: Foundations of the ETA Certified AI Empowered Professional. A quick refresher:
- Step 1: Assess
Is this a good task for AI, and what's at stake?
- Step 2: Direct
Brief the AI clearly: goal, context, constraints, done criteria.
- Step 3: Assure
Check the output before you rely on it.
- Step 4: Protect
Keep people, data and your organization safe.
- Step 5: Tune
Iterate, then save what works for next time.
Use the steps in order on every task, and loop back to Tune when results fall short.
ADAPT is ETA's AI fluency framework: Assess whether and how to use AI for a task, Direct it with a clear brief, Assure the output by checking it, Protect people and data, and Tune results over time.
What Claude can see
Claude only works with what is in front of it. That is the single most useful thing to understand.
| Where you work | What Claude can see |
|---|---|
| A new chat | Your messages, files you attach, and anything it looks up if web search is on |
| A Project | The Project's instructions and knowledge files, plus the chat |
| A connector | The app data your account can reach, when you ask Claude to use it |
| Memory (if your plan and settings have it on) | Short notes it keeps from past chats |
Claude doesn't see your email, drive or calendar unless a connector is set up and you, or your admin, have allowed it. It also doesn't know your organization's unwritten rules. If something matters, put it in the brief.
Your plan decides your data rules
Claude comes in personal plans (Free, Pro and Max) and business plans (Team and Enterprise). Business plans add single sign-on, admin controls and shared Projects, and Anthropic states that it doesn't train its models on business plan content by default. Personal plans have their own privacy settings that you control.
Before you paste anything from work, answer three questions:
- Is Claude approved here? Check your organization's AI policy or ask IT.
- Which plan am I on? A personal account is not the same as your company's Team or Enterprise plan, even if you log in with your work email.
- What data is OK? Use your organization's OK, ask first and never rules.
Passwords, access codes, security keys and other credentials
- Personal information
Social Security numbers, birth dates, home addresses, personal phone numbers
- Health information
Diagnoses, medical leave details, patient records
- Financial account details
Bank and card numbers, individual pay records
- Student records
Grades, discipline and other student information
- Confidential business and client information
Unreleased results, pricing, contracts, legal matters, anything under a confidentiality agreement
Some organizations approve specific tools for some of this data. That decision belongs to your organization, not to you. If you're unsure, treat it as off-limits and ask. You rarely need real details: redact, substitute or generalize.
Workplace example: a nonprofit grants team
Marisol manages grants at a regional food bank. Her team has a Claude Team plan. She uses it to draft first versions of grant narratives from the food bank's annual report and program data. She never pastes donor names or client records: the policy lists them as never. And she checks every figure in the draft against the annual report, because a wrong number in a grant application is hard to take back.
Assess: is this a good Claude task?
Good fits are tasks where a draft saves time and you can check the result: summaries of documents you have, first drafts, rewriting for a new audience, analysis of data you can verify, brainstorming options. Poor fits are tasks where nobody can check the answer, where the data is off limits, or where a person's judgment is the whole point (a hiring decision, a clinical or legal call).
Try it
Open Claude (or your organization's approved assistant if you don't have Claude). Ask it to explain its own limits:
I'm about to use you for work. In plain language, tell me:
1. What you can see in this chat right now, and what you can't.
2. Whether you can browse the web in this chat.
3. Three kinds of work tasks you're well suited to, and three where a person should decide.
Keep it under 200 words.Then check the answer against your organization's settings or IT guidance. If anything differs, trust your organization's settings.
Key takeaways
- Claude only sees what's in the chat, the Project, its memory and the connectors you allow.
- Check whether Claude is approved, which plan you're on, and which data is OK before you share work content.
- Good Claude tasks produce a draft or analysis you can check; judgment calls stay with people.