The Claude AI Tutorial That Starts With Work Already on Your Desk

Most tutorials about Claude or any AI tool follow the same frustrating arc: fifteen minutes of background on large language models, a few screenshots of the interface, an example where someone asks it to write a haiku, and then a vague conclusion about "the possibilities." You close the tab and go back to your actual work, no better equipped than before.

This is not that tutorial.

What follows starts with four scenarios you probably recognize from your own desk — a client call that needs a recap, a report you haven't had time to read, a variance memo for the board, and a position you need to stress-test before a hard conversation. For each one, I'll walk through the prompt structure, what you're trying to get, and what you need to verify before you use the output. No toy examples. No warm-up exercises. Just the work.

If you've read the foundational overview of how to use Claude, this is where you take those principles into real situations. If you haven't done anything with Claude yet, The Leverage Starter is designed to get you from zero to a working session in under two hours — starting with tasks exactly like these.

Why the Prompt Structure Matters More Than the Prompt

Before the scenarios: a brief principle you'll use in all four. Claude does not read your mind. It reads what you give it. When output is weak, the first diagnosis is almost always the prompt — not the model. Specifically, three things degrade output faster than anything else: vague context, missing role-framing, and no stated output format.

Role-framing means telling Claude what perspective to take. "You are drafting this for a senior partner reviewing a complex tax position" produces materially different output than an unframed request. Context means giving Claude enough background to understand the situation — who the client is, what the subject matter is, what constraints apply. Output format means being explicit about what you want back: a one-page memo, a three-paragraph email, a structured analysis with headers, a list of objections. When you specify all three, Claude can allocate its processing to getting the substance right instead of guessing at the container.

The scenarios below build all three into every prompt structure. You'll notice the pattern quickly.

Scenario 1: The Post-Call Recap Email

You just finished a 90-minute call with a client — a board member, a principal, a wealth advisory client, whoever. You have rough notes. They're in whatever form your notes always take: incomplete sentences, jargon, a few action items buried in the middle, and a follow-up question you forgot to ask. You need to send a professional recap within the hour.

The Prompt Structure

Open with role-framing: "You are drafting a professional recap email from a senior advisor to a client following a strategy call." Then provide context in a dense paragraph — not bullet points, not cleaned-up notes, just the actual notes pasted in. Include the client name (first name only is fine, or a pseudonym — never upload identifying information that isn't necessary), the general subject of the call, the key topics covered, and any action items or open questions you flagged. Then specify output: "Draft a recap email of no more than 350 words. Use a professional but direct tone. Structure it as: two-sentence summary of what was discussed, key decisions or conclusions reached, action items with owner and timeline, and one sentence on next steps."

That structure — role, context, format — will get you 80% of the way to a usable draft in a single pass. The gap between "80% there" and "ready to send" is where your judgment goes. Read it for factual accuracy. Verify that action items are attributed correctly. Make sure the tone matches your relationship with this specific client. Claude does not know whether your client prefers formal salutations or whether there's a history that changes the framing — you do.

What to Verify Before Sending

Check every factual claim, every number, every attributed statement. Claude will occasionally consolidate two separate points into one or soften a commitment that was actually firm. Read it as you would any draft before it goes out under your name, because it is going out under your name.

Scenario 2: The Report You Haven't Had Time to Read

You have a 40-page report — an appraisal, a due diligence package, a government filing, an opposing counsel's expert report — and a client meeting in two hours. You have not read it. You need to walk into that meeting with command of the material.

The Prompt Structure

Paste the full text of the document, or as much of it as is relevant. (Claude has a large context window; most documents under 100 pages paste cleanly.) Then prompt: "You are a senior professional reviewing this document before a client meeting. Summarize the key findings, conclusions, and any material risks or open issues. Structure the output as: executive summary (150 words), key points by section (three to five bullets per major section), and a separate section flagging anything that appears inconsistent, uncertain, or that warrants follow-up questions." Then add: "Do not soften the findings. If there is bad news in this document, state it plainly."

That last instruction matters more than it might seem. Claude tends toward diplomatic language. If you're walking into a negotiation or a client conversation where you need to know the worst-case read of a document, you have to ask for it explicitly.

The prompt patterns for professionals post goes deeper on how to structure document-review prompts for different document types — regulatory filings, contracts, financial statements, and expert reports each have slightly different structures worth using.

What to Verify Before the Meeting

This is not a substitute for reading the document. It is a way to arrive at the meeting with a map of the territory. In the meeting, you may need to find specific language quickly; having read the summary means you know where to look. After the meeting, read the sections that came up. The summary is your pre-briefing, not your permanent record.

Also: never paste documents containing client PII, tax identification numbers, account numbers, or anything that would be a confidentiality problem outside your walls. If the document has that material, redact it first or work from a sanitized version. The Fiduciary Firewall briefing covers the specific rules you need to understand before putting any client material into any AI tool.

Scenario 3: The Variance Memo for the Board

You need to explain a variance — revenue came in 18% below forecast, a project ran 30% over budget, a fund returned 600 basis points below the benchmark. The board wants a memo. You know exactly what happened. What you don't have is the time or inclination to write three drafts of an explanation that will be read by people who will immediately probe the weakest part of whatever you write.

The Prompt Structure

Role-frame for the reader, not the writer: "You are drafting a variance explanation memo for a board of directors. The audience is sophisticated and will probe the explanation for gaps. The tone should be direct and analytical — no hedging, no passive voice, no burying the lead." Then provide the factual context: what the metric was, what the target was, what the actual result was, the primary drivers of the variance, any secondary factors, and what the plan is going forward. Then specify: "Draft a memo of 400–500 words. Lead with the variance and the number. Explain causes in order of magnitude. Close with corrective actions and timeline. Do not start with background or context — start with what happened."

What you get back is usually a structurally sound draft that has organized your explanation better than a cold-start draft would. The substance is yours. Claude is imposing structure and economy on it. That is valuable.

The Second Pass

Read the draft and ask yourself: "Where will the board push back?" Then run a second prompt: "Given this memo, what are the three most likely follow-up questions a board member would ask, and where are the weakest points in the current explanation?" This is Claude functioning as a pressure-test, not a ghostwriter. The questions it surfaces are usually the right questions — because it's reading the same text the board will read and it's not emotionally attached to the explanation you've given.

If you work in finance, the Leveraged CPA and Finance Professional course covers exactly this workflow — structured writing for senior financial audiences — with more depth than this post allows.

Scenario 4: Stress-Testing Your Position Before a Hard Conversation

You are going into a deposition prep session, a negotiation, a client conversation where the other side has a strong position, or a board meeting where you expect opposition. You have your position. You need someone to argue against it — specifically and hard — before you walk into the room.

This is where Claude earns its keep in ways that surprised most of the professionals I've introduced it to. It is a very capable adversarial thinker when you ask it to be one.

The Prompt Structure

"You are an experienced opposing counsel / opposing analyst / skeptical board member [choose the role]. I am going to describe my position on [subject]. Your job is to identify every weakness in my argument, every fact I might be wrong about, every assumption I'm making that could be challenged, and the strongest version of the counterargument. Do not be diplomatic. Do not validate my position first. Go directly to the critique." Then state your position in full — as you would present it in the room.

What comes back will be uncomfortable if your position has real gaps. That is the point. Better to find the gap here than in the room. You can then follow up: "Now steelman my position — what are the strongest responses to the objections you raised?" and "What information would I need to find or verify to close the weakest points in my argument?"

This is a thinking-partner workflow, not a drafting workflow. You are not publishing the output. You are using it to sharpen your own thinking before a situation where weak thinking has consequences. Attorneys preparing for depositions, deal professionals walking into term sheet negotiations, consultants presenting an uncomfortable recommendation — this is the scenario where Claude's lack of emotional investment in your position is a feature.

The deeper guide to using Claude at work covers the thinking-partner mode in more detail, including how to structure multi-turn conversations where you're building an argument iteratively rather than in a single prompt.

The Workflow Across All Four Scenarios

The pattern is consistent enough to be worth naming explicitly. Every effective professional use of Claude follows the same structure: role-frame for the appropriate reader or perspective, provide dense factual context (not cleaned up, just accurate), specify the output format explicitly, and verify the output against your own knowledge before using it. That's the whole tutorial. The scenarios above are just the pattern applied to real situations.

What changes across scenarios is what you're optimizing for. In the recap email, you're optimizing for professional tone and factual accuracy. In the document summary, you're optimizing for completeness and flag-raising. In the variance memo, you're optimizing for structural clarity and defensibility. In the stress-test, you're optimizing for adversarial completeness. The prompt structure shifts slightly with each, but the underlying discipline is identical.

The other thing worth saying: none of this requires you to become a "prompt engineer." That phrase has accumulated a mystique it does not deserve. What you are actually doing is communicating clearly about what you need and who you're speaking to. Senior professionals do this every day when they brief a junior associate, instruct a paralegal, or hand off a task to a capable analyst. Claude is a capable analyst. Brief it accordingly.

If you want to develop this into a systematic practice — not just occasional use but an integrated part of how you work — the memo-writing workflow post covers the writing scenarios in more depth, and The Leverage Starter will take you through your first structured Claude session with real work from your own desk. It's two hours. You'll leave with three or four outputs you can actually use, and a prompt pattern library you'll keep using afterward.

The professionals who get the most out of Claude are not the ones who spent the most time reading about it. They're the ones who started with the work they already had and built the habit from there.

Frequently Asked Questions

What are the three elements that degrade Claude AI output the most?

The three things that degrade Claude output faster than anything else are vague context, missing role-framing, and no stated output format. When you specify all three — role, context, and output format — Claude can allocate its processing to getting the substance right instead of guessing at the container.

How should I prompt Claude to write a post-call recap email?

Open with role-framing such as "You are drafting a professional recap email from a senior advisor to a client following a strategy call." Then paste your raw notes as dense context, and specify the output format: no more than 350 words, a professional but direct tone, structured as a two-sentence summary, key decisions, action items with owner and timeline, and one sentence on next steps. This structure gets you 80% of the way to a usable draft in a single pass.

Can Claude summarize a long report before a client meeting?

Yes. Paste the full text of the document — Claude has a large context window and most documents under 100 pages paste cleanly — then ask for an executive summary, key points by section, and a separate section flagging anything inconsistent or uncertain. Add the instruction "Do not soften the findings" because Claude tends toward diplomatic language by default, and you need the worst-case read if you are walking into a negotiation or client conversation.

How do I use Claude to pressure-test my position before a negotiation or deposition?

Role-frame Claude as an opposing counsel, opposing analyst, or skeptical board member, then instruct it to identify every weakness in your argument, every fact you might be wrong about, and the strongest version of the counterargument — without being diplomatic. After receiving the critique, follow up by asking Claude to steelman your position and identify what information you would need to close the weakest points. This is a thinking-partner workflow, not a drafting workflow.

What should I verify before using Claude's output in a professional context?

Check every factual claim, every number, and every attributed statement before using Claude's output. Claude will occasionally consolidate two separate points into one or soften a commitment that was actually firm. Read it as you would any draft before it goes out under your name — because it is going out under your name.

Do I need to become a prompt engineer to use Claude effectively at work?

No. The phrase "prompt engineer" has accumulated a mystique it does not deserve. What you are actually doing is communicating clearly about what you need and who you are speaking to — which senior professionals already do every day when briefing a junior associate or handing off a task to a capable analyst. Claude is a capable analyst; brief it accordingly.

---

Where this goes next

Ready to put this to work? See The Leverage Starter — or The Leveraged Consultant if you want the broader path.

Related reading from The Briefing

• How to Use Claude AI: A Practical Guide for Senior Professionals
• Using Claude AI for Work: The Daily Operating Model for Senior Professionals
• The Claude AI Prompts That Senior Professionals Keep Using Every Week

Not sure which program fits where you are? take the 2-minute course-fit quiz, or browse the full course catalog.