Skip to content
SkillCurioPractical paths into AI
← All practical guides

Start here · IT professionals

Where should an IT professional start with AI?

Start with a task you already understand. Use your experience to judge whether AI makes the result more useful.

If you work in desktop support, infrastructure, or IT operations, you already bring something useful to AI: you know what a real problem looks like. You know when an instruction is incomplete, when an answer needs checking, and when a change needs approval.

You can begin by applying that judgment to a small exercise. You do not need to decide on a new job title or buy a course first.

1. Choose a task you can evaluate

Pick something familiar enough that you can spot mistakes. A troubleshooting checklist, a clearer explanation for a user, or a draft knowledge-base article is a useful starting point.

Keep the exercise small. “Help me understand AI” is hard to finish. “Draft a checklist for a fictional second-monitor problem, using only these observations” gives you something concrete to inspect.

Use an AI tool your organization permits. For this exercise, invent the scenario and every detail. Keep real tickets, screenshots, client information, credentials, device identifiers, and internal documents out of it. Follow your employer’s rules before using workplace information or trying a suggested change.

2. Try a 20-minute exercise

Use this fictional scenario: a laptop’s second monitor is blank after reconnecting a dock. The laptop screen works. You have not checked the cable, monitor input, or display settings.

Give your permitted AI tool this prompt:

Help me draft a troubleshooting checklist for this fictional scenario: a laptop’s second monitor is blank after reconnecting a dock. The laptop screen works. The cable, monitor input, and display settings have not been checked. Start with observations and low-impact checks. Ask for missing details rather than inventing them. Separate confirmed facts from possible causes. Flag anything requiring admin rights, a restart, software installation, or a configuration change for review. Do not provide commands or scripts. Use plain English and explain why each check matters.

Read the draft as if a new colleague had written it. Ask the tool to revise one weak section, then compare the two versions. Keep your own notes about what changed and why.

3. Judge the answer before using it

A confident answer still needs review. Use these questions:

  • Did it invent a dock model, operating-system version, or cause?
  • Are the checks in a sensible order, with assumptions made visible?
  • Could a user follow the wording without guessing?
  • Does it distinguish observation from changes that need approval?
  • What would you confirm against the actual device documentation?

If it suggests changing a setting or taking an action you cannot explain, stop at the draft. Check the relevant vendor documentation and your organization’s procedures before proceeding.

For broader context, NIST’s AI Risk Management Framework addresses trustworthiness in the design, use, and evaluation of AI. You can explore that framework as your work moves beyond individual practice.

4. Make your first week manageable

  1. Session one: choose the fictional task and write down what a good answer should include.
  2. Session two: generate a draft, mark mistakes and assumptions, and request a revision.
  3. Session three: write a short before-and-after note: what improved, what needed correction, and whether the review effort was worthwhile.

Three short sessions are a suggested pace, not a deadline. The useful result is a better understanding of where the tool helps and where your judgment is essential.

5. Choose your next learning step

If the terminology feels unfamiliar, start with AI Fundamentals. If you want clearer instructions and more reliable drafts, explore Prompt Engineering. Once you can explain and review a repeatable workflow, consider No-Code AI Automation.

These are starting options, not a required sequence. The Find my path quiz suggests resources based on your goal and experience level.

Your next question can be simple: “What is one part of my work I understand well enough to practice improving?” Start there, and keep the result small enough to check.

Keep learning

Get SkillCurio Dispatch in your inbox.

The complete issue, every two weeks: useful AI developments, what they mean for your skills, and a practical next step. Unsubscribe anytime.

Delivered via Buttondown. Your address is never shared or used for anything else.