Effective AI use begins before you type a single word into a prompt window. It begins with a professional judgment call that most people never make consciously — and that omission is exactly why so much AI-assisted work falls short of professional standards.
The five questions below are not a checklist to follow mechanically. They are the mental habit of a competent AI professional. Run through them quickly at first. Over time they become second nature.
1. What is the actual output I need?
This sounds obvious. It is not. Most people open an AI tool with a vague sense of what they want — "help me write this email" or "summarize this report" — without specifying what a successful result actually looks like.
Before you prompt, define the output with professional precision. What format? What length? What audience? What tone? What level of technical detail? The clarity you bring to this question directly determines the quality of what you get back.
A project manager asking for "a summary" will get something generic. A project manager asking for "a 5-bullet executive summary of this document, written for a CFO who has not read the background materials, focused on cost implications and timeline risks" will get something useful.
2. How much does it matter if this is wrong?
This is the risk assessment question, and it is the most important one on the list. The answer shapes everything that follows — how much scrutiny you apply, whether you need human review, and whether AI is the right tool at all.
Low-stakes tasks — an internal draft, a brainstorm list, a first-pass outline — can tolerate more error. The consequences of a mistake are low, and the efficiency gains are high. Use AI aggressively here.
High-stakes tasks — a client-facing analysis, a compliance document, a performance review, any communication that carries your professional signature — demand a different standard. AI can still help, but the verification discipline required is proportionally higher.
3. Do I have the expertise to evaluate the output?
This question is the one professionals most frequently skip, and it is the one that produces the most embarrassing failures. AI output can be fluent, confident, and completely wrong. If you cannot evaluate the output, you cannot use it responsibly.
A marketing professional who asks AI to draft campaign copy can evaluate it — they know what good looks like. The same professional asking AI to analyze statistical significance in A/B test results may not be able to spot an error in the reasoning.
Know the boundary of your own expertise. When you are operating near that boundary, increase your verification effort accordingly. When you are clearly beyond it, get a qualified human review.
4. What context does AI need that it does not have?
AI operates only on what you give it. It does not know your organization's tone of voice, your manager's preferences, the political sensitivity of a situation, or the history behind a project. The gap between what AI knows and what the task actually requires is where most AI output fails to meet professional standards.
Before prompting, inventory the context gap. What does this task require that AI cannot infer from a generic prompt? Then provide it. This is not extra work — it is the difference between output you can use and output you have to rewrite from scratch.
5. Is AI actually the right tool for this?
Sometimes the answer is no. Not because AI is incapable, but because the task requires something AI cannot provide: your specific institutional knowledge, your relationship with the reader, your professional judgment about a nuanced situation, or simply the credibility that comes from your name on something you wrote yourself.
The professional who uses AI for everything is not more efficient. They are outsourcing their judgment. The professional who knows when to reach for AI and when to do the work themselves — that is the person who builds a durable reputation in an AI-enabled world.
These five questions take less than a minute once they are internalized. The investment in building this habit will pay dividends across every AI-assisted task you ever take on.
Built on the J.E.T. Model
This article is part of the Judgment pillar of the J.E.T. framework — a professional competency model for AI use. Explore it in full in Don't Wait.