10 Copy-Paste Prompts to Audit Your Job for AI Automation Risk

Last updated: October 17, 2026

“Will AI take my job?” is the question everyone asks. Workday’s answer, launched at its Workday Rising 2026 conference, is that you are asking the wrong question. The right question is which tasks inside your job AI will take first — because AI rarely changes a whole role at once. It changes tasks, and the skills that matter shift with them.

Quick answer: Workday's Work Intelligence scores AI risk task by task, not job by job. Copy these 10 prompts to audit your own role — and build your upskill plan tonight.

Workday’s product only serves enterprise customers, though. So here is the same exercise as ten copy-paste prompts you can run tonight with any AI assistant, plus the scorecard for reading the results.

What Workday just launched

Work Intelligence was announced on 13 October 2026 at Workday Rising 2026 in Las Vegas. It draws on Workday Skills Cloud — used by more than 3,450 customers and built on an ontology of over 75,000 skills — and breaks each job into the tasks people actually do. Each task lands in one of three buckets: can be done wholly with AI, partially with AI, or without AI.

“The basic building block of work is shifting from the job to the task — and without that view, most companies are left guessing whether to hire, build skills, automate, or redesign roles, with no way to know which guess is right.” — Nikki Helmer, vice president of talent management and insights, Workday

Workday’s 2026 Global Workforce Report, cited at the launch, found that 76% of heavy AI users expect AI to create new career opportunities. Where Workday gives leaders data to redesign roles, the prompts below give you the data to redesign yourself.

Before you run the prompts: gather these three things

The 10 copy-paste prompts

1. Build the task inventory

Act as a workforce analyst. Here is my role and everything I did in the last two weeks:

[PASTE YOUR TASK LIST]

Organise these into a clean task inventory. Group near-duplicates into single tasks, split any task that actually contains two different kinds of work, and number them. For each task, add: what I do, what tools or systems it touches, and who else is involved.

2. Time-weight every task

For each numbered task in the inventory, assign an estimated weekly time in hours. Where I have not given times, estimate from the task type and flag those rows as estimated. Then sort the list by hours, descending, and show the running percentage of my total week.

A task AI could take matters little at 1% of your week; a partially automatable one at 20% is where your quarter gets decided.

3. Catch the hidden tasks

Ask me five targeted questions that would reveal work I do regularly but did not list — things like waiting on approvals, fixing other people's errors, context-switching between tools, onboarding, emotional labour, and informal mentoring. Then update the inventory with anything I confirm.

The tasks we forget to mention are often the most human ones — and the hardest to automate.

4. Score each task for automation risk

For each task in my inventory, score its AI automation risk from 1 to 5:
1 = AI cannot touch this (needs physical presence, trust, or real authority)
2 = AI can assist but a human must drive
3 = AI can do most of it with human review
4 = AI can do it fully with spot checks
5 = already automatable with tools available today

Use three inputs for each score: how repetitive and rule-based the task is, how much judgment under uncertainty it needs, and how costly a mistake would be. Show your reasoning in one line per task, and give me the weighted average using my time percentages.

Wholly, partially, or not at all — rendered as a number you can compare across tasks.

5. Separate what dies from what moves

Split my tasks into three lists, Workday-style: tasks AI will likely take over entirely, tasks where AI becomes a co-worker, and tasks that must stay human. For the "take over" list, estimate the year range when tooling could do it reliably. For the "co-worker" list, estimate the hours I could reclaim weekly if I used AI well.

Dying tasks need replacing; co-worker tasks need tooling. Most career anxiety comes from treating them the same way.

6. Test for judgment, not just repetition

For my top five time-weighted tasks, describe the hardest judgment call inside each one: the decision I made where there was no rule to follow. Then assess whether current AI systems could make that call with the context they would have access to — and what would need to change for the answer to become yes.

7. See it through your employer’s eyes

Pretend you are my employer's finance team in 2027. Rank my tasks by the ratio of cost savings from automating them to the risk of automating them. Which three tasks would you target first, and what would stop you from automating each one?

Your risk is not what AI can do — it is what AI can do cheaply and safely. This reveals which tasks are targets, not just candidates.

8. Map the skills that survive

Based on my task scores, list the skills I already use that become MORE valuable as my routine tasks get automated: judgment, relationships, domain expertise, taste, crisis handling, negotiation. For each, name the task in my inventory where I demonstrate it today, and rate my current strength as beginner, solid, or expert.

As Workday puts it, AI changes tasks, and as tasks change, the skills that matter shift with them. You want to hold the ones that appreciate.

9. Write the 90-day upskill plan

Using my risk scores, skill strengths, and time savings, write a 90-day plan: one skill to deepen (my best AI-proof asset), one task to automate for myself (reclaiming hours), and one new capability to build from my 'co-worker' list. Include weekly actions, a free or low-cost way to learn each one, and a concrete milestone to prove progress at day 30, 60 and 90.

Audit without action is just anxiety with a spreadsheet; three commitments is a plan you can keep.

10. Set up the quarterly re-check

Create a short re-audit checklist I can run every three months: which three questions to re-ask, which new AI capabilities to check against my task list, and which leading indicators (job postings in my field, tool launches, my employer's tech choices) should make me re-score a task early.

AI capability moves in jumps — the scores you generate tonight will drift, and the re-check keeps the audit honest.

How to read your scorecard

Once prompt 5 is done, drop your tasks into this three-bucket table:

BucketWhat it meansYour move
Automate (scores 4–5)AI can do this fully, or already canLearn to supervise it rather than compete with it — or accept the hours will shrink
Augment (scores 2–3)AI does the heavy lifting, you keep the judgmentAdopt the tools yourself; this is where the reclaimed hours fund your upskilling
Keep human (score 1)Needs trust, authority, presence, or tasteDouble down — these are your career moat and your negotiation leverage

A healthy result: a shrinking “automate” column of busywork, a fat “augment” middle where you get faster, and a “keep human” column you can name in an interview. If you are applying for new roles too, our AI Job Hunt OS: 7-Day Plan to Land Interviews pairs well with this audit — it is built for the other side of the same problem.

The honest limits of this exercise

Frequently asked questions

Is this the same thing as Workday’s Work Intelligence?

No — it is a personal approximation of the same logic. Work Intelligence is an enterprise product that draws on company HR and skills data; these prompts apply the same task-level thinking using only what you know about your own work. Same philosophy, completely different data and scale.

Could running this audit make me look replaceable?

Done privately, it does the opposite: it shows you exactly where your human judgment lives, which is the part no automation plan can touch. Most people discover their “keep human” column is larger than their fear suggested — and now it is documented.

Which AI assistant should I use for the prompts?

Any general-purpose assistant that accepts long pasted text works — ChatGPT, Claude, Gemini, Meta AI, or Copilot. For sharper results, run the scoring prompts twice in different tools and compare; where they agree, the signal is stronger. None of the prompts need a paid tier.

How often should I re-run the audit?

Quarterly is the right cadence. AI capabilities move in step-changes, not smooth curves, so your scores will sit still for months and then jump when a new model or agent framework lands. Prompt 10 gives you the re-check ritual.

Does this work for creative or manual jobs too?

Yes — the task logic is universal. A video editor’s “sync audio” scores differently from their “client taste calls”; a warehouse lead’s “rostering” differs from “de-escalating a dispute”. Swap the examples; the prompts handle the rest. When you are ready to delegate the automatable tasks to agents, our 8 copy-paste skill packs show how agent skills structure that handoff.

Sources and methodology

Sources: Workday launches Work Intelligence AI offering to help make automation decisions (The Hindu, 14 Oct 2026) — the launch report this article is based on; Workday launches Work Intelligence to show how AI changes job tasks (Unite.AI, 13 Oct 2026) — task-level framing and early-access detail; Workday launches Work Intelligence to help companies prepare jobs for AI (Analytics India Magazine, 15 Oct 2026) — Skills Cloud figures and Workday Rising context. Related reading on this site: AI Job Hunt OS: 7-Day Plan to Land Interviews, Free Meta Muse Skills: 8 Copy-Paste Skill Packs.

Method: claims rest on Workday’s 13–14 October 2026 launch materials as reported by the sources above; vendor claims (Skills Cloud figures, 76% workforce-report stat) are attributed to Workday and reported, not independently verified. The ten prompts and scorecard are original to this article, desk-written on 16 October 2026 and not hands-on validated against real roles. Verification date: 16 October 2026. This is a career-education article, not financial or employment advice.

Have a burning question about this topic?
Feel free to email us at contact@openaimaster.ai — we are happy to help!