@itsmariahbrunner — AI at Work
Your AI risk score tells you where your role is exposed. This guide tells you exactly what to do about it — including a prompt that maps out your entire role and builds your handoff plan.
The real question isn't your score
A high AI risk score doesn't mean you're getting fired. It means a significant portion of your current job is doable by a machine — which is actually useful information if you act on it. Most people see the score and panic. The people who get ahead see it as a task list.
The goal isn't to protect every part of your role. It's to figure out which parts you should hand off as fast as possible so you can pour your time and energy into the parts that make you genuinely irreplaceable. That shift is what separates people who thrive in the AI era from people who get left behind.
This guide gives you a framework for making that decision — and a prompt that does the analysis for your specific role, not a generic one.
The two-question framework for every task in your job
Pattern-based tasks follow rules, templates, or repeatable steps — writing a similar email every week, summarizing information, formatting documents, pulling data. Judgment-based tasks require reading context, making calls with incomplete information, weighing competing priorities. Pattern = hand off. Judgment = protect.
Some tasks are low-stakes if Claude gets it slightly wrong — a first draft that needs editing, a summary you'll verify. Others carry real risk if the output is off — a message to a difficult client, a strategic recommendation, a sensitive personnel decision. High stakes = you stay in the loop. Low stakes = Claude drafts, you review.
Pattern-based and low stakes. First drafts, templates, research summaries, meeting recaps, repetitive communications, formatting, status updates. Claude produces, you do a quick review and send.
Judgment-based but something Claude can help you think through — a difficult conversation you need to prepare for, a strategy document, a decision with tradeoffs. You lead. Claude adds structure, stress-tests your thinking, helps you see angles you missed.
The goal isn't to replace yourself. It's to free yourself from the work that has nothing to do with why you're actually valuable.
The handoff decision — by task type
I want you to help me build a complete task audit and handoff plan for my role. My role: [your job title] What I actually do day to day: [write 3–5 sentences describing your real daily work — the more specific, the better this analysis will be] My industry/company type: [e.g. marketing agency, corporate finance, small business owner, healthcare admin] Step 1 — Map my full task landscape List every category of work my role typically involves. For each, give me a rough estimate of how much of my week it consumes (high / medium / low). Don't just list job description tasks — think about the actual work: emails, prep, coordination, documentation, meetings, thinking time, reactive work. Step 2 — Score each category on two dimensions For every category you listed, rate it on: — Pattern vs. judgment (1 = almost entirely pattern-based, 5 = almost entirely judgment-based) — Stakes if wrong (1 = low stakes, easily corrected, 5 = high stakes, real consequences) Present this as a simple table. Step 3 — Build my handoff tiers Based on your scoring, sort my tasks into three groups: TIER 1 — Hand off now: Low judgment + low stakes. These are eating time I should be spending elsewhere. For each, give me the exact prompt or workflow I should use in Claude. TIER 2 — Use Claude as a co-pilot: Higher judgment but Claude can add real value in prep, structure, or drafting. For each, describe specifically what I use Claude for and where I stay in control. TIER 3 — Keep fully human: These require my experience, my relationships, or my accountability. Claude can assist at the edges but the work is mine. Step 4 — My 30-day starting point Given everything above, tell me: — The single highest-leverage task I should start handing off this week (and why) — The one skill I should develop in the next 30 days to get the most value from handing things off — What I should stop feeling guilty about delegating to AI Be specific to my role. Generic advice about AI isn't useful — I want an analysis that actually reflects what I do.
When you hand off the pattern-based work, you get time back. The question is what you do with it. These are the areas where investing your freed-up capacity pays off most.
The kind of specific, hard-earned knowledge that takes years to develop. AI can retrieve information — it can't replicate what you know from having done the work for a decade.
Who calls you when something goes wrong. Who vouches for you. Who wants to work with you specifically. These compound over time in a way AI cannot replicate.
Making good calls when the information is incomplete, the stakes are real, and the situation is novel. This is what separates senior people from everyone else — and AI isn't close.
Knowing what good looks like. Having taste. Being able to direct, evaluate, and elevate — not just produce. Claude can generate. You decide what's worth keeping.
Motivating someone who's struggling. Giving feedback that actually lands. Building a team that trusts each other. This is irreducibly human — and undervalued by most people until AI makes it obvious.
Reading the room. Understanding the organizational dynamics behind a decision. Knowing what's actually going on beyond what's written in a document. AI works with what it's given. You work with everything you know.
Want to go further?
Follow me for daily content on how to actually use Claude at work — not theory, not hype, just practical workflows for real jobs.