AI for the Workplace with Claude: Complete Online Course Guide

If you've been using AI at work without a system behind it — asking ChatGPT a question here, trying Copilot there — you've probably noticed the ceiling. Results are inconsistent. You re-prompt three times to get what you want. You can't remember what worked last Tuesday, and nothing carries forward. Noble Desktop's AI for the Workplace with Claude course is built for exactly that problem. It's a 2-day, hands-on training where you go from zero Claude experience to a fully configured, personalized AI workflow you use on your own files, in your own job, the day after class.

This guide walks through the full curriculum — every module, every capability, every capstone exercise — and includes four free lesson previews from instructor Brian Simms so you can see the tone and substance before you enroll. By the end of this page you'll know whether this is the right course for your role and what system you'll walk out with.

Meet the Course

This isn't a lecture series on how AI works under the hood. It's a working session. In Brian's own words from the opening module:

"This is not just going to be a lecture about AI theory. This is going to be a working course. In just a few modules, you will have used Claude to write, analyze, summarize, research, and create real deliverables. By the end, you'll have a complete system built around your own work."

Watch the full introduction below. Each video on this page has a searchable, clickable transcript — jump to any line to skip straight to the point you care about.

What You Will Learn

The course is organized around building a personalized Claude system — not just a list of features. You will learn how to:

• Write structured prompts that produce accurate results for writing, analysis, summarization, and research
• Set up and use Projects, Artifacts, Styles, and Memory to create a personalized Claude workflow
• Configure Claude Cowork — the desktop agent — with appropriate guardrails, file permissions, and network settings
• Build a personal context system (about-me, brand-voice, working-style) that improves with every session
• Connect Claude to the tools you already use: Slack, Gmail, Google Drive, Notion, Excel, and PowerPoint
• Install and use role-specific plugins with slash commands for domain-specific workflows
• Set up scheduled tasks that automate recurring work — weekly reports, daily briefings, Friday status summaries
• Apply Claude across job functions: data, marketing, HR, operations, project management, and finance

The agenda packs a lot into two days. Day 1 covers claude.ai — the browser product you can open today. Day 2 covers Claude Cowork — the desktop agent that reads and writes files on your computer — plus connectors, Office add-ins, plugins, and scheduled tasks. You leave with a capstone project: a working Claude system configured for your actual job.

Getting Started: The Right Mental Model

The first module is about building a solid mental model of what Claude actually is — and what makes it different from the other AI tools you may have used. Brian is upfront about this on day one:

"My goal in this course is not to convince you that Claude is the best AI tool in the world, though it's pretty awesome. It's to give you an honest, accurate understanding of what it does, what it doesn't do well, and how you can use it more effectively. By the end, you'll be equipped to make your own informed decisions about when it's time to use Claude, or maybe it's time to use something else."

That honest framing runs through the whole course. Module 6 does a direct, head-to-head comparison of Claude against ChatGPT, Copilot, and Gemini — and tells you when to reach for each one.

What Claude Actually Is

Claude is an AI assistant built by Anthropic, a San Francisco-based AI safety company founded in 2021 by former members of OpenAI. You interact with it in plain English — it writes, summarizes, codes, brainstorms, and reasons through complex problems. You can use it through claude.ai (the web interface), a desktop application (critical for Cowork), or an API for developers.

Two things are worth understanding up front, because they shape how you'll use it for the rest of your career:

1. Claude is not a search engine. When you ask a question, it's not going out to Google and pulling back results — it generates responses based on patterns learned during training. That means it can be confidently wrong. This is called hallucination, and the course addresses it head-on.
2. The current generation is Claude 4.5. Launched in 2023, the platform has already been through several generations of models. You'll learn when to use each tier — Haiku for speed, Sonnet for daily work, Opus for maximum reasoning depth.

Here's Brian walking through exactly that framing — it's about 90 seconds and sets the foundation for everything that follows.

The Anatomy of a Great Prompt

The single biggest factor in how useful Claude is for you isn't the model, isn't the plan you're on, and isn't any specific feature — it's how well you communicate with it. Module 2 drills the anatomy of an effective prompt until it's automatic: context, task, format, constraints, examples. Five parts. Every prompt.

The payoff is enormous. The same question, phrased two different ways, produces two wildly different outputs. Weak prompts get you generic paragraphs you have to rework. Strong prompts get you something close to the final answer on the first try.

Five core use cases you'll practice

The course uses five recurring prompt patterns — one for each high-value job on most professionals' plates. Every pattern follows the same anatomy, just tuned for the use case.

Day 1 Curriculum: claude.ai Mastery

Day 1 is seven modules and runs roughly 6.5 hours with a 60-minute lunch break. Every module ends with a hands-on exercise. By the end of the day you've used Claude to draft, analyze, and summarize real documents; built your first Project; customized a Style; set a Memory; and compared Claude's output against another AI tool you already use.

Module 1 — Getting Started with Claude AI (60 min)

What Claude is, how it works in plain language, the context window (up to 1,000,000 tokens — about ten Harry Potter books of working memory), and how to pick a model tier. You'll take your first pass at the claude.ai interface, send your first real prompt, and follow up with one iteration.

Module 2 — Prompting Fundamentals and Working with Files (90 min)

The single biggest factor in how useful Claude is for you isn't the model or the plan — it's how well you communicate with it. You'll learn the anatomy of an effective prompt (context, task, format, constraints, examples), the five core use cases (writing, summarization, analysis, research, creative problem-solving), and how to work with uploaded files — images, PDFs, Word docs, spreadsheets, screenshots.

Module 3 — Projects: Organizing Ongoing Work (75 min)

In a regular Claude conversation, every chat starts from zero. Projects eliminate that. A Project is a dedicated workspace where you set context once and every conversation inside it inherits that context. You'll create a real Project — custom instructions, uploaded reference docs, style rules — and use it on a real task from your work.

Module 4 — Artifacts, Styles, and Memory (45 min)

Artifacts are standalone deliverables — documents, comparison tables, spreadsheets, code, visualizations — that Claude produces alongside the conversation. Styles let you encode your voice (concise, data-driven, no jargon). Memory lets Claude retain what it learns about you across conversations. Together these three features are the difference between using Claude and using a generic chatbot.

Module 5 — Intermediate Prompting and Advanced Features (75 min)

Role-based prompting ("Act as a budget analyst reviewing this report"). Chain-of-thought reasoning ("walk me through your reasoning before making a recommendation"). Multi-step prompting that breaks a big task into three to five verified steps. Web search grounding. Extended Thinking for deep reasoning. And the technique Brian says he would teach if he could only teach one thing: the "You Ask Me" pattern, where you hand Claude the goal and ask it to interview you for context before it starts.

Module 6 — Claude vs. Other AI Tools (60 min)

An honest framework for choosing between Claude, ChatGPT, Copilot, and Gemini. Start with where your data lives (Microsoft 365 → Copilot; Google Workspace → Gemini; neither → Claude or ChatGPT). Then what task you're doing (writing and analysis → Claude; image generation → ChatGPT; spreadsheet data → Copilot or Gemini). Then your organization's constraints. There is no single best AI tool — you'll leave with a decision tree you can use forever.

Module 7 — Day 1 Wrap-Up and Day 2 Preparation (15 min)

What to install, which folder to prepare, which accounts to authenticate — so Day 2 starts with everyone ready to go.

Projects, Artifacts, Styles, and Memory

These four features are what separate "using a chatbot" from "having a working AI system." In a plain Claude conversation, every chat starts from zero. You re-explain your role, your project, your preferences, every single time. Modules 3 and 4 eliminate that tax for good.

A Project is a dedicated workspace. You set context once — who you are, what you're working on, how you want Claude to respond — and every conversation inside it inherits that setup automatically. Artifacts are the standalone deliverables Claude produces. Styles encode your voice. Memory is the persistent thread that carries across conversations. Together, they turn Claude from a clever stranger into a briefed colleague.

Day 2 Curriculum: Cowork, Automation, and Your System

Day 2 is where Claude goes from "a really smart person you could call on the phone" to "a really smart person sitting at the desk next to you, with access to your file cabinet, your email, and your calendar." Eight modules covering Claude Cowork, your personal context system, guardrails, connectors, plugins, scheduled tasks, industry-specific applications, and a capstone that ties everything together.

Module 8 — Introduction to Claude Cowork (45 min)

Cowork is the desktop AI agent. It reads and writes files on your actual file system, plans multi-step tasks, executes autonomously, queues and runs tasks in parallel, and asks clarifying questions instead of guessing. You'll set up Cowork, point it at a prepared folder, and run your first multi-step task.

Module 9 — Building Your Context System (60 min)

Context files are the real force multiplier. Three foundational files do most of the work. Plus Global Instructions that apply to every session (always ask clarifying questions before starting; default to markdown; use active voice; if you're unsure about a fact, say so — do not guess). This module is where most students report the biggest "aha" moment.

Module 10 — Guardrails, Permissions, and Safe Use (45 min)

File system permissions (dedicated working folder, never point Cowork at your entire Documents directory). Network egress modes (off, package managers only, all domains). Data privacy — including a critical warning that Cowork conversation history is stored locally on your computer and is NOT captured in Audit Logs, the Compliance API, or Data Exports, so you should not use it for regulated workloads. And prompt injection: what it is, why it matters, and how to defend against it.

Module 11 — Connectors, Claude in Chrome, and Office Add-Ins (60 min)

Link Claude to external tools via MCP — the Model Context Protocol — and query them live from conversation. Covered: Slack, Google Drive, Gmail, Notion, Figma, Asana, Linear. Plus Claude in Chrome (browser-based research and data extraction, with "ask before acting" safeguards), and the native Excel and PowerPoint add-ins that read entire workbooks, write formulas, create pivot tables, and turn spreadsheet analysis into presentation-ready decks.

Module 12 — Plugins and Scheduled Tasks (60 min)

Plugins transform Claude from a generalist into a domain expert. Each one bundles Skills (specialized knowledge), Commands (slash-command workflows like /analyze and /report), Connectors, and Sub-Agents. Scheduled Tasks (launched February 2026) run automatically on a schedule — "every Friday at 3pm, create a weekly status summary from my project files and save it as weekly-summary-[date].md in my Reports folder." You'll install a plugin matched to your role and schedule your first recurring task.

Module 13 — Industry Applications (60 min)

You pick at least two scenarios and work through them end-to-end:

• Data and Analysis — upload a dataset, identify trends, generate visualizations, produce a summary report
• Marketing and Communications — campaign copy, content calendar, social media variations
• Human Resources — position description analysis, job posting creation, classification standards
• Operations and Administration — Excel restructuring, pivot tables, document library cleanup
• Project Management — timelines, status reports, risk assessments
• Finance — budget analysis, sensitivity tables, financial summaries

Module 14 — Capstone: Build Your Personal Claude System (75 min)

Seven steps that produce a working Claude system you take home and use the next day. Details below.

Module 15 — Wrap-Up and Next Steps (15 min)

A Week 1 Action Plan — one real task per day, refining as you go — plus an honest look at remaining limitations.

Your Context System: The Real Force Multiplier

If there's one thing that separates people who "use Claude sometimes" from people who run a durable AI-assisted workflow, it's this: a small, living set of context files that describe who they are and how they want to work. Claude reads them at the start of every session. They compound in value the more you use them.

The course teaches three foundational files. Start small — five to ten sentences each — and expand as you learn what Claude asks you for twice.

The Capstone Method

The capstone is the final exercise of Day 2, and it's the reason most students say they'd take the class again. In 75 minutes you walk through seven concrete steps that produce your own, running Claude system — configured for your job, connected to your tools, with at least one recurring task already scheduled.

The steps look almost trivial on paper. The magic is that each one uses Claude to design the next one. Step 2 — "ask Claude which of your tasks it can help with most" — is where the whole system starts writing itself.

Your Week 1 Action Plan

The course ends with a light-touch plan for the week after class. One real task per day, refining as you go. It's designed to make sure what you built doesn't sit unused — the most common failure mode for any training course.

Honest Limitations

One of the things students most appreciate about the course is that Brian is upfront about what Claude still doesn't do well. You leave with an accurate mental model — not a marketing pitch — and a clearer sense of when to use Claude versus when to reach for something else.

Key Techniques You'll Take With You

Every course promises you'll "learn AI." The difference with this one is specific, named techniques that keep paying off long after the class ends. A few of them:

The anatomy of an effective prompt. Context, task, format, constraints, examples. Weak prompt: "Write a summary of our project." Strong prompt: "I'm a PM at [org]. Summarize our Q3 website redesign project for my director. Include timeline status, budget variance, and two key risks. Keep it under 250 words. Use a professional, direct tone." Brian's rule of thumb: if you find yourself re-prompting Claude three or four times to get what you want, the problem usually isn't Claude — it's the original prompt. Spend the extra 30 seconds upfront.

The "You Ask Me" technique. Instead of crafting a perfect prompt, tell Claude the goal and ask it to interview you first. "I want to draft a performance review for a team member so that I can prepare for a conversation with them next week. Don't start yet. Ask me clarifying questions first." This is the technique Brian would teach if he could only teach one.

Managing your usage. Toggle web search off when you don't need current information — a single 10,000-word article pulled into context consumes a massive chunk of your quota. Start new conversations when switching topics. Use Projects to store reference docs. Choose the right model tier for the task.

Chain-of-thought reasoning. "Walk me through your reasoning before making a recommendation." Produces more accurate results on complex tasks — and makes it easier to spot where Claude's logic might be wrong.

Who Should Enroll

This course is built for professionals who already use a computer for most of their work and want Claude to be a reliable, ongoing part of how they do it. No programming experience required. No AI background required. The learners who get the most out of it tend to be:

• Knowledge workers in operations, marketing, HR, finance, or project management
• Analysts and researchers who work with documents and data
• Team leads who want their team to use AI responsibly and consistently
• Individual contributors tired of the "start-from-scratch" tax on every AI conversation
• Anyone who has tried ChatGPT casually and wants to build something more durable

What you need before Day 2: a paid Claude subscription (Pro at minimum, Max recommended), the Claude desktop app installed, a folder with three to five real work files prepared for exercises, and — optionally — credentials ready for any of Slack, Google Drive, or Gmail you'd like to connect.

What You Leave With

You don't leave this course with a certificate of completion and a vague sense of "AI literacy." You leave with a running system: a personalized Claude setup configured for your actual job, context files that describe who you are and how you work, connectors wired to the tools you already use, at least one scheduled task producing output every week, and a week-one action plan for locking it in.

Ready to enroll? Noble Desktop offers this course as a live, expert-led 2-day class online, with a free retake within one year if you want to come back after you've had a chance to use what you built. See upcoming dates and pricing on the AI for the Workplace with Claude course page, or browse more AI training in Noble Desktop's Generative AI Certificate and Data Science & AI Certificate programs.