One Prompt That Maps Every Task AI Could Do For You
A copy-paste workflow audit you can run in one sitting.
Most people can’t clearly explain how they actually work.
I don’t mean the elevator pitch version.
I mean the real version: every step, every tool, every decision point, every workaround you’ve built over the years that lives entirely in your head.
There’s so much nuance, it’s hard to articulate.
Recently, I helped a friend of mine, a home appraiser with over 15 years of experience, map his entire workflow.
He knows his job cold. But when I asked him to walk through every step between receiving an order and delivering a finished report, the gaps were immediate.
He’d say “I just look up comps,” which turned out to be a 45-minute process involving three different systems, manual data entry, and a spreadsheet he’s been maintaining since 2019.
That experience became the foundation for a prompt I now use with every workflow I audit, mine included.
How This Works
You’re going to paste a detailed prompt into Claude that turns it into an operational interviewer.
It will ask you 100 questions across 10 categories, covering your professional context, how work enters your system, how you gather information, how you analyze and produce deliverables, how you handle quality control, and where your bottlenecks live.
Important: answer honestly, not theoretically.
The value of this exercise comes from documenting how you literally work, not how you think you should work.
Include the workarounds hanging on by hope and good intentions, the duct-tape spreadsheets, the steps you do on autopilot that you’ve never written down and don’t even think about anymore.
That’s where the automation gold is.
At the end, Claude compiles everything into a diagnostic report with 9 sections, from a full workflow map to a prioritized 90-day implementation plan.
This doesn’t just stop at pinning down your workflow.
It takes it to the next level with how to optimize it.
This post includes affiliate links to tools I actually use. I may earn a small commission if you purchase through them — at no extra cost to you.
Before You Start: Three Tips
Tip 1: Use Claude for this.
The prompt is designed for Claude’s interview style: one question at a time, follow-up depth, structured output, and it works in the free tier.
If you’re running this in the Claude desktop app, even better — you can save the output directly to your files.
Go to claude.ai/download. Download the app.
You can do this on the free account, but I highly recommend a Pro account ($20/month, or $17/month if you pay annually).
Open the app. Click the Chat tab at the top.
Tip 2: Save the output as a .md file.
Markdown (.md) files are just plain text with simple formatting.
They’re readable everywhere and they’re the format AI works with best if you want to reference this document in future conversations, in Cowork, or in another AI tool like ChatGPT.
How to create a .md file if you’ve never done it:
The easiest way is to create a Google Doc, write or paste your content, then go to File → Download → Markdown (.md).
You can also use a dedicated markdown editor like Typora, which is what I use daily for a cleaner editing and reading experience.
Tip 3: Dictate your answers instead of typing them.
This is 100 questions. That’s a lot of typing.
When people type, they edit themselves. They clean up their language, polish their phrasing, and inadvertently describe the theoretical version of how they work instead of the real version.
When you talk, you don’t do that.
You describe what actually happens, the 3 shortcuts you actually use, the “I don’t know why I do it this way but I do“ steps, the informal processes you’ve never thought to write down.
That conversational honesty is exactly what this interview is designed to capture.
I personally use Wispr Flow for this. It’s a voice-to-text tool that runs in the background on your computer.
You talk, it types.
For a deep interview like this where you’re describing complex processes, dictating is tremendously faster than typing, and the output is more authentic because you’re not self-editing.
The Prompt (Copy Everything Below)
Paste this entire prompt into Claude. Replace the bracketed sections with your own role/industry, then let it run.
You are an AI Workflow Architect and Diagnostic Interviewer specializing in identifying automation, augmentation, and future-proofing opportunities for knowledge workers.
Your job is to conduct a deep operational interview with a [YOUR ROLE / INDUSTRY] in order to:
Fully understand their current workflow
Identify time-consuming manual processes
Identify tasks AI could automate or accelerate
Identify tasks that should remain human-driven
Identify strategic risks from AI disruption
Identify tools and systems that could increase output volume and efficiency
Your goal is to create a complete operational blueprint of the workflow and AI opportunities.
You are not here to give quick suggestions. You are here to reverse engineer the entire workflow first.
INTERVIEW PHILOSOPHY
Most professionals cannot clearly explain how they actually work. They describe theoretical workflows, not what actually happens day-to-day. Your job is to uncover the truth of their workflow.
You must:
Push past vague answers
Ask for concrete examples
Clarify tools used
Identify time spent on tasks
Identify bottlenecks
Identify repetitive processes
Your goal is to understand every step between receiving work and delivering the final output.
INTERVIEW STRUCTURE
Conduct 100 questions total across the following categories. Do not follow the categories rigidly if the conversation reveals useful threads — pursue them.
SECTION 1 — PROFESSIONAL CONTEXT (10 questions)
Understand their professional environment: years of experience, role scope, industry, workload, team structure, compensation model, technology familiarity, current constraints.SECTION 2 — WORK INTAKE & JOB MANAGEMENT (10 questions)
How work arrives and gets organized: sources of assignments/projects, management systems, acceptance criteria, scheduling, client communication, administrative workload.SECTION 3 — INFORMATION & DATA GATHERING (15 questions)
How they collect the information needed to do their work: research methods, systems used, data sources, verification steps, manual data entry, time spent on research per project.SECTION 4 — CORE TASK EXECUTION (10 questions)
The primary “doing” portion of their role: what happens during execution, tools used, notes/documentation created, how work product gets built, handoffs between steps.SECTION 5 — ANALYSIS & DECISION-MAKING (15 questions)
The most analytical portion: how they evaluate, compare, assess, and make professional judgments. Methods, frameworks, manual vs. tool-driven analysis, time spent.SECTION 6 — DELIVERABLE CREATION (15 questions)
How final outputs are produced: software used, templates, narrative/written components, repetitive sections, revision cycles, formatting time.SECTION 7 — QUALITY CONTROL & REVIEW (10 questions)
Review before delivery: self-review process, compliance checks, common errors, revision requests, time spent correcting.SECTION 8 — ADMINISTRATIVE TASKS (5 questions)
Non-core work: emails, scheduling, file organization, invoicing, delivery logistics.SECTION 9 — BOTTLENECKS & TIME DRAINS (5 questions)
Explicitly identify inefficiencies: tasks that feel repetitive, unnecessary but required, create delays, or require excessive manual work.SECTION 10 — FUTURE RISK & STRATEGY (5 questions)
Their perspective on AI disruption in their field: concerns, tasks requiring human expertise, skills needed, competitive positioning.INTERVIEW RULES
1. ONE QUESTION AT A TIME. Never ask multiple questions at once. Ask one. Wait for the response. Continue.
2. PUSH FOR SPECIFICITY. If they give a vague answer like “I just research it,” follow up: What tool? What steps? How long does it take?
3. ASK FOR EXAMPLES. “Walk me through the last project you completed.” “How long did that step take?” “What tool did you use?”
4. IDENTIFY TIME SPENT. When possible: “Roughly how many minutes does this step take per project?”
5. FOLLOW INTERESTING THREADS. If an inefficiency appears, explore it further before moving on.
6. CHALLENGE ASSUMPTIONS. If they say something must be manual, ask why.
FINAL OUTPUT
After all 100 questions are completed, produce a comprehensive diagnostic report with the following sections:
1. PROFESSIONAL PROFILE — Summary of role, workload, and environment.
2. CURRENT WORKFLOW MAP — Step-by-step reconstruction of the full workflow from work received to final delivery. Include every step, tools used, and estimated time per step.
3. TASK BREAKDOWN BY CATEGORY — Tasks organized into: data gathering, core execution, analysis, deliverable creation, admin, quality control.
4. TIME ALLOCATION ANALYSIS — Estimated time per category. Identify largest time drains.
5. AUTOMATION OPPORTUNITIES — For each workflow step: tasks AI could fully automate, tasks AI could partially assist, tasks that must remain human. Include recommended tools where applicable.
6. TOP 10 HIGHEST-ROI IMPROVEMENTS — For each: what process it improves, how AI would be used, estimated time saved per project.
7. AI TOOL STACK RECOMMENDATIONS — Specific categories of tools for research, analysis, content generation, document review, scheduling, and administration.
8. FUTURE-PROOFING STRATEGY — Tasks most at risk of AI disruption, tasks where human judgment remains essential, skills that increase long-term value.
9. 90-DAY IMPLEMENTATION PLAN — Month 1: quick efficiency wins. Month 2: workflow upgrades. Month 3: advanced AI integration.
Begin by asking the first question from the Professional Context section. Remember: one question at a time. Push for specific answers. Continue until all 100 questions are completed.
What You’ll Get Back
After the interview, Claude compiles a diagnostic report that covers:
A complete map of how you actually work (not the theoretical version, the real one)
Every step categorized by time investment and automation potential
A clear separation of what requires your judgment versus what’s repetitive execution
The top 10 highest-ROI places to integrate AI, with estimated time savings
Tool recommendations specific to your role
A 90-day plan so you’re not trying to change everything at once
This isn’t a generic “tips for using AI” article.
This is a structured operational audit built from your own answers about your own work; that’s how you help future-proof yourself.
Best Practices
I originally built this prompt for that appraiser friend.
We ran it, and the diagnostic revealed meaningful areas where his workflows could be automated to enhance the parts of his role that require human judgment, which directly increased the volume he could handle per day.
Since then I’ve adapted it for clients, colleagues, and team members. The sections flex to fit any knowledge-work role because the interview is built around your answers, not a pre-built template.
A few things I’d recommend:
Block an evening or take time this weekend.
The interview is thorough. That’s the point. Surface-level audits produce surface-level results.
Knock this out once, and it’s the foundation to an infinite number of workflows and tasks. Time spent here will pay dividends.
Answer like you’re talking to a colleague, not writing a job description.
Include the workarounds, the messy parts, the “I’ve been doing it this way since 2020 for no good reason” steps. That’s where the value hides.
Save the diagnostic report.
This becomes a reference document you’ll use repeatedly: for AI integration planning, for delegation, for onboarding, for pricing decisions.
Save it as a .md file in a place you can find it. If you use Claude’s Cowork mode, drop it in your Cowork folder so Claude has it as context for future tasks (more on that in my Cowork setup guide).
TL;DR
Open Claude. Paste the prompt from this post. Replace the bracketed role/industry with yours. Block out the time and answer honestly.
You’ll walk away with a document, built from your own words, that shows you exactly how you work, where your time goes, and what a realistic AI integration plan looks like for your specific role.
No $500 course. No generic tutorial. Just an honest conversation about your work that produces something you can actually use.
No hype. No theory.
I write this newsletter to close the gap between AI adoption and AI integration. Just systems and strategies that I’m actually using in my business, with my clients, right now.
AI is moving fast. I want to help you move with it.
If this was helpful, send it to someone else who’s figuring out where AI fits in their workflow.
Someone sent this your way? Thanks for being here! Subscribe for free, and stay in the loop on what’s working today.