Tips for Better Results

You’ve probably noticed that sometimes AI gives you exactly what you need on the first try, and other times it produces something completely off-base. The difference usually isn’t the AI. It’s how you asked. Getting good results from AI is a skill, and like any skill, it gets better with practice.

This guide covers the principles and techniques that consistently lead to better AI interactions. These aren’t technical tricks. They’re practical habits that anyone can pick up. Think of it as the difference between telling a new employee “handle that thing” versus giving them clear, specific instructions. Same person, wildly different results.

How to give clear, specific instructions

The single biggest factor in the quality of AI output is the quality of your input. Vague prompts produce vague results. Specific prompts produce useful results.

Include the five key elements

When you’re asking AI for something substantial, try to include:

What you want: the task itself
Who it’s for: the audience
The format: how you want the result structured
The tone: formal, casual, technical, friendly
Constraints: length limits, things to include or avoid

Here’s the difference in practice:

Without the five elements:

“Write something about our new product.”

With all five elements:

“Write a product announcement email (1) for our existing customers (2) as a short email with a subject line and three paragraphs (3) in an enthusiastic but not over-the-top tone (4) that’s under 200 words and doesn’t mention pricing since that’s not finalized yet (5).”

The second prompt will produce something usable on the first try. The first will need several rounds of back and forth.

Give context about your situation

AI doesn’t know anything about you unless you tell it. The more relevant context you share, the more tailored the result:

“I’m a middle school teacher who needs to explain climate change to 12-year-olds. Many of my students come from families that are skeptical about the topic, so I need to be factual and respectful without being preachy.”

That context changes everything: vocabulary level, tone, sensitivity, framing. Without it, you’d get a generic explanation that might not work in your classroom.

Show, don’t just tell

When possible, give the AI an example of what you’re looking for:

“Write a bio for our company website. Here’s one from a similar company that has the tone I want: [paste example]. I want something similar in style but obviously about our company. Here are the facts to include: [your information].”

“I need a subject line for a marketing email. Here are three subject lines that performed well for us in the past: [examples]. Write 10 new subject lines in a similar style for our upcoming summer sale.”

Examples are worth a thousand words of description. If you have something close to what you want, sharing it will get you to the right result faster than any amount of detailed instruction.

Be explicit about what you don’t want

Sometimes it’s easier to define what you want by ruling out what you don’t:

“Write a company bio that’s professional but not corporate-speak. Don’t use phrases like ‘synergy,’ ‘leverage,’ ‘best-in-class,’ or ‘thought leader.’ Don’t start with ‘Founded in…’ and find a more interesting opening.”

“Suggest 10 names for my new coffee shop. Don’t use puns. Don’t use the words ‘bean,’ ‘brew,’ or ‘cup’ since they’re overused. I want names that feel warm and inviting, not trendy or hipster.”

The art of iterating

Here’s something that separates people who love AI tools from people who gave up on them: the first response is rarely the final answer. The real skill is in the follow-up.

Refine, don’t restart

When the AI gives you something close but not quite right, resist the urge to start over. Build on what you have:

“That’s really close, but the second paragraph is too long and the ending feels abrupt. Shorten paragraph two to about three sentences and add a warmer closing.”

“I like the structure of this draft but the tone is too casual for a business context. Keep the structure exactly as is but make the language more polished and professional.”

“The first half of this is great — keep it. The second half goes off track. It should focus on practical advice, not theory.”

This works better than writing a whole new prompt because the AI keeps the context of what it already produced and your earlier instructions. Be specific about what’s working, not just what needs to change. Saying “the opening is strong, but the conclusion just repeats the introduction” helps the AI understand what to keep and what to revise.

Ask for multiple options

When you’re not sure exactly what you want, ask for variety:

“Give me three different versions of this email — one that’s warm and personal, one that’s crisp and professional, and one that’s somewhere in between.”

“Write five different opening paragraphs for this blog post. Vary the approach: try starting with a question, a statistic, an anecdote, a bold claim, and a scene-setting description.”

Looking at multiple options is often faster than trying to describe the perfect version in advance. You’ll know it when you see it.

Build up piece by piece

For complex projects, don’t try to get everything right at once. Work on one element at a time:

First, get the structure right: “Give me an outline for this presentation.”
Then develop each section: “Now flesh out section two with specific examples.”
Then refine the tone: “This section sounds too academic. Make it more accessible.”
Finally, polish: “Read through the whole thing and fix any inconsistencies or rough transitions.”

Each step builds on the last, and you can give focused feedback at each stage.

Understanding AI limitations

Using AI well means understanding where it’s strong and where it falls short. This isn’t pessimism. It’s just about using the right tool for the job.

What AI does well

Generating ideas and options: brainstorming, suggesting alternatives, exploring possibilities
Structuring and organizing information: turning messy notes into clear outlines, creating comparisons
Adapting tone and style: making writing more formal, casual, concise, or detailed
Explaining concepts: breaking down complex ideas at different levels
First drafts: getting past the blank page quickly
Summarizing: distilling long documents into key points
Template-style tasks: anything with a clear structure or repeatable pattern

Where AI struggles

Current events and real-time information: unless web search is turned on, its knowledge has a cutoff date
Precise numerical calculations: it makes arithmetic errors, especially with complex math
Deep domain expertise: it knows a little about a lot, which sometimes means it misses nuances an expert would catch
Your personal context: it doesn’t know your company culture, your family dynamics, or your history unless you explain it
Predicting the future: it can analyze trends but can’t tell you what will actually happen
Being concise on the first try: AI tends toward verbosity and often needs to be told to trim down

The confidence trap

AI almost always sounds confident, even when it’s wrong. It doesn’t say “I’m not sure about this” as often as it should. This is one of the most important things to understand.

A response that reads as authoritative and well-structured can still contain errors. How fluent the writing sounds has nothing to do with how accurate the information is. Keep this in mind, especially with facts, figures, and claims.

When to trust AI output vs. verify independently

Not everything needs to be fact-checked, but some things absolutely do. Here’s a practical framework.

Trust more (but still review)

Writing style and structure: if it reads well and says what you mean, it’s fine
General creative ideas: brainstorming output, topic suggestions, structural options
Well-known facts: widely established information that’s unlikely to be wrong
Formatting and organization: how information is presented

Verify always

Specific numbers and statistics: double-check any data point that matters to your decision
Claims about specific people, companies, or events: AI sometimes makes up details that sound plausible
Legal, medical, or financial advice: always confirm with a qualified professional
Dates and timelines: especially for recent events
Technical specifications: model numbers, product features, pricing
Quotes and citations: AI sometimes generates fake quotes or cites papers that don’t exist

A simple rule of thumb

Ask yourself: “If this information is wrong, what’s the worst that happens?” If the answer is “my email is slightly awkward,” you probably don’t need to verify. If the answer is “I make a bad financial decision” or “I give someone incorrect medical information,” verify independently.

Common mistakes and how to avoid them

Mistake 1: Being too vague

Instead of: “Help me with my resume.” Try: “Review my resume for a senior marketing manager position. I want to emphasize my experience with digital campaigns and team leadership. The company is a mid-size tech startup with a casual culture. Here’s my current resume: [paste it].”

Mistake 2: Accepting the first response

The first response is a starting point, not a finished product. Always read it critically and ask for changes. Even a simple “make this shorter and punchier” can make a big difference.

Mistake 3: Not providing examples

If you have a specific style, tone, or format in mind, show the AI an example rather than trying to describe it. “Write it like this example” almost always works better than “write it in a warm but professional tone with short paragraphs.”

Mistake 4: Asking for too much at once

A prompt that asks for an email, a social media post, a presentation outline, and talking points all at once will produce mediocre results for each. Focus on one task at a time and give each one proper attention.

Mistake 5: Not mentioning your audience

The same information should be presented very differently to a CEO than to a new intern, to a customer than to a colleague, to a child than to an adult. Always mention who will be reading or hearing the output.

Mistake 6: Forgetting to mention constraints

Length limits, things to avoid, required elements, deadlines, format requirements. If there are rules your output needs to follow, mention them upfront rather than hoping the AI guesses right.

Mistake 7: Using AI for everything

AI is useful for many things but it’s not always the right tool. Sometimes a quick Google search is faster. Sometimes picking up the phone is more effective. Sometimes you just need to sit and think without a machine generating ideas for you. Use AI where it genuinely helps, not everywhere.

Building your AI skills over time

The best way to improve is to practice with intention. After each AI interaction, briefly consider:

Did I get a useful result? If not, what could I have said differently?
Did I iterate, or did I just accept the first response?
Was there context I forgot to include?

Over time, you’ll develop an intuition for how to communicate effectively with AI, and that intuition transfers across different tools and platforms.

For practical applications of these principles, explore our other guides: Writing and Communication for email and document help, Research and Analysis for information gathering, Organization and Productivity for managing your time and tasks, and Creative Projects for brainstorming and content creation.