Most of us use AI the same way. We type a command, get a mediocre answer, wonder why we bothered. But a small shift in how you ask can make a surprisingly big difference. Here's how to start getting answers that are actually worth your time.
------
Last week you asked AI to help you write a client email.
What came back was stiff, generic, and honestly kind of useless. You spent 20 minutes rewriting the whole thing and thought, “why did I even bother?”
You’re not alone. An Upwork Research Institute study found that 77% of employees using AI say it’s actually added to their workload, and nearly half don’t know how to get the results their employers expect from it.
Here’s the thing: the tool usually isn’t the problem. It’s what we’re asking it to do.
A small shift in approach makes a real difference, and it doesn’t require any technical skill. Just a different starting point.
Most of us use AI the same way. “Write me an email.” “Draft a report.” “Summarize this.” We treat it like a content machine, and then wonder why the content isn’t very good.
But the people getting real value from AI? They’re not using it to produce finished work.
They’re using it to think things through before they act.
The highest-performing organizations aren’t the ones with the best AI tools. They’re the ones where people have conversations with AI, going back and forth, rather than firing off one-line commands.
That’s not our opinion. Microsoft found exactly this in their 2025 Work Trend Index, based on data from 31,000 workers worldwide.
Here’s what that looks like in a regular workday:
|
Brainstorming |
“I’ve got a staffing problem on a project and I’m stuck. Give me 10 ways to approach it—include a few I probably haven’t thought of.” |
|
Pressure-testing |
“Here’s my plan for this rollout. Poke holes in it. What are the biggest risks I’m missing?” |
|
Meeting prep |
“I’ve got a meeting with a client tomorrow about renewing their contract. What should I be ready for?” |
|
Summarizing |
“Here’s a 20-page safety report. I need the 5 key points for a quick conversation with my director.” |
|
Simplifying |
“We’re rolling out a new expense policy. Help me explain it to the team in plain language.” |
See the difference? None of these ask AI to do your job. They ask AI to help you do your job better. That’s where the real time savings come from.
One real example: an operations leader at an AI education company shared publicly how they upload draft board presentations into AI every quarter and ask it to challenge their thinking, once as a growth-focused board member, once as a risk-cautious one.
The result wasn’t a better slide deck. It was a better-prepared leader walking into the room.
When you do want AI to write something (an email, a summary, a first draft) the quality of what comes back depends almost entirely on what you put in. Here’s a quick example:
Look at the “After” column. Notice what’s different? There’s context (who you are and what the situation is), constraints (the tone you want, what to include), and it’s set up for iteration (you could easily follow up with “make it shorter” or “add a line apologizing for the inconvenience.”).
That’s really all there is to it. Give AI enough to work with, tell it what you don’t want, and treat the first answer as a starting point, not the final product.
Research from the OECD and Stanford’s AI Index backs this up: people see real productivity gains of 10–25% on everyday tasks like writing and research, but only when they know how to work with the tool.
The technology isn’t the bottleneck. Knowing how to ask is.
A quick reality check: AI still gets things wrong. It can sound confident while being completely off base. Your judgment matters. Always review what it gives you. But when you ask better questions, you spend less time fixing bad output and more time building on good output. That’s the real win.
You don’t need to become a prompt engineer. You don’t need a training course.
Just pick one thing this week:
an email
a meeting you need to prepare for
a long document you’ve been putting off reading
...and try being more specific with what you ask. Tell AI who you are, what you need, and what good looks like.
You might be surprised how much better the answers get when you change the question.
---------
Techify helps organizations get more from the technology they already use. If your team wants to go deeper on AI, get in touch.