How do I know when to use AI vs. do it myself?

How Do I Know When to Use AI vs. Do It Myself?

by Stephen Wu, a retired engineering manager, spent several years building AI workflows for his teams and now trying to make AI something everyone can understand and use successfully

Knowing when AI will actually save you time—versus when it'll just add extra steps—takes practice. The decision isn't always obvious, and using AI when you shouldn't can slow you down rather than help. Here's how to think through when AI makes sense.

The Basic Question

Ask yourself: "Will explaining this to AI, reviewing its output, and refining the result take less time and energy than just doing it myself?"

Sometimes the answer is clearly yes. Sometimes it's clearly no. Often it's somewhere in between, and you'll learn through experience.

When AI Usually Helps

AI tends to be useful when the task involves generating options, processing information, or creating a first draft of something you'll refine.

Use AI when:

  • You need multiple approaches to a problem and want to see different options
  • You're starting from scratch and a rough draft would help you get going
  • You're working with a lot of information that needs organizing or summarizing
  • You need something explained in a different way than you've encountered
  • The task is routine enough that you can quickly spot if AI gets it wrong
  • You have time to review and refine what AI produces

AI prompt: "I need to write an email declining a meeting invitation without offending anyone. Can you give me three different approaches I could take, from most direct to most diplomatic?"

In this case, AI helps by showing you different communication strategies. You pick the one that fits your relationship with the person and adjust the phrasing. AI saves you time thinking through different approaches.

When Doing It Yourself Is Faster

AI adds steps to your workflow: formulating a prompt, waiting for a response, evaluating whether it's useful, refining if needed. If a task is quick and straightforward for you, those extra steps aren't worth it.

Don't use AI when:

  • You can do it in two minutes without thinking much
  • You already know exactly what you want to say or do
  • The task requires personal knowledge or context that would take longer to explain than to just do the work
  • Getting AI's output right would require so much back-and-forth that you'd save time doing it yourself
  • The task is so critical that you'd agonize over whether AI got it right

Example: You need to reply to a coworker's quick question about meeting at 2pm.

AI prompt: "How should I respond to confirm I can meet at 2pm?"

This is overkill. Just write "2pm works for me, see you then." The task is simpler than the AI interaction would be.

The Learning Curve Factor

Early on, using AI takes longer because you're still figuring out how to ask questions and evaluate responses. As you get more practiced, the decision point shifts—tasks that initially seemed faster to do yourself might become faster with AI once you know how to use it efficiently.

First time using AI: You spend 10 minutes trying to get AI to help you draft an email, go through three rounds of revisions, and end up writing most of it yourself anyway.

After practice: You quickly prompt AI for a structure, adapt the parts that work, ignore the parts that don't, and have a polished draft in 3 minutes.

The same task becomes more AI-appropriate as your skill with AI improves.

When the Stakes Matter

High stakes change the calculation. If something is really important, you might use AI even though it takes more time, because having a second perspective or seeing multiple approaches is worth the extra effort.

Low stakes: Quick Slack message to your team about lunch plans. Just write it.

High stakes: Email to your boss explaining why a project is behind schedule. Use AI to help you think through how to frame the situation, what information to include, and how to position next steps. Then carefully refine it in your own voice.

The importance of the outcome justifies the additional time and careful review.

A Practical Decision Framework

Here's a quick mental checklist:

Lean toward AI if:

  • The task would take you more than 5 minutes
  • You're stuck or not sure how to start
  • You'd benefit from seeing multiple approaches
  • It's something you do rarely and aren't practiced at
  • Getting it wrong has consequences worth avoiding

Lean toward doing it yourself if:

  • You can finish it in under 2 minutes
  • You know exactly what you want
  • It requires explaining a lot of context
  • It's personal or sensitive in ways AI wouldn't understand
  • You'd spend more time checking AI's work than doing it yourself

Real Example: Writing a Meeting Agenda

Scenario: You need to create an agenda for tomorrow's team meeting.

Consider AI if: You're not sure what topics should be included, or you want to see how someone else might structure it, or you're dealing with multiple competing priorities and want help organizing them.

AI prompt: "I'm running a team meeting tomorrow. We need to discuss the Q4 project timeline, address some communication issues that came up last week, and plan for an upcoming client presentation. Can you suggest an agenda structure that addresses these topics in a logical order?"

AI helps you think through sequencing and might suggest things you hadn't considered, like starting with quick wins before tackling harder topics.

Skip AI if: You run these meetings every week, you know exactly what needs to be covered and in what order, and writing the agenda takes you 90 seconds.

Learning Your Own Patterns

Pay attention to which tasks end up being faster with AI and which ones slow you down. You'll develop intuition about your personal threshold.

Some people find AI helpful for almost all writing tasks. Others only use it for specific types of problems. There's no universal right answer—it depends on your working style, your skill with AI, and the specific task.

The Bottom Line

The decision to use AI should make your work easier, not add busywork. If you find yourself using AI out of habit rather than because it's genuinely helpful, step back and ask whether it's actually saving you time.

Conversely, if you're avoiding AI for everything, you might be missing opportunities where 30 seconds with AI could save you 30 minutes of work.

The skill is knowing the difference.