Design principles for choosing and using AI tools

When I was seven, my dad took me with him on a grocery run.

He was a delivery driver for one of New Zealand's largest supermarket chains and had a reputation for spotting a bargain. We drove across town early on a Saturday morning to a smaller store he knew had better prices. He walked the aisles with a list and a plan, showing me what to look for.

In case your dad didn't teach you how to grocery shop, allow me to share two nuggets of gold he taught me that day.

First, the cheapest products are often on the lowest shelves, out of sight.

Second, the bright yellow "special" tags didn't always mean that product is on special. Lift the tag and check the price underneath. Often it reveals a $0.01 saving, making you buy a product you don't need.

I can't remember the specific products we bought that morning, but I still use the principles when I'm shopping: have a plan, make a list, look for opportunities, and check you're getting a good deal.

I think about that now when I see advice about AI.

A lot of it focuses on which tools to use. But I've found it far more useful to develop a set of design principles that help me decide what is appropriate for me to use, at a given time.

From fear to intentional use#

My own journey with AI didn't start with curiosity. It started with fear.

At the end of 2022, while on maternity leave, I saw ChatGPT in action for the first time. My immediate reaction was to wonder whether my job would still exist when I returned to work.

That became the catalyst for learning everything I could about generative AI. I wanted to understand how these systems worked, how I could use them, and what skills I needed to develop to stay relevant.

I focused first on human capabilities:

• Conversation
• Critical thinking
• Storytelling

Over time, I added:

• Delegation
• Strategic vision
• Connection with people

That shift guides how I use AI today. I don't start with tools. I start with curiosity and capability.

The problem with most AI advice#

Much of the current conversation about AI lacks precision.

We talk about it as if it's a single, all-powerful entity:

• "AI helped me do that"
• "Just use AI for this"
• "AI is taking all the jobs"

This isn't specific enough to be useful.

AI didn't help you get the job. You might have used Copilot to map your CV to a job description, used ChatGPT to rephrase your experience, or used Claude to prepare for the questions you might be asked in the interview.

That level of specificity helps you understand what worked, what didn't, and where to improve. Ask for it from others and make it a habit to be specific when you're talking about how you use AI.

There's also a tendency toward over-optimism.

It's easy to be drawn in by simple promises. A new tool claims it can manage your calendar, summarise your thinking, or organise your life. In exchange, it asks for access to your emails, your contacts, your personal data.

We rarely pause to consider what we are handing over.

This is where data literacy becomes essential. And data literacy is foundational to AI literacy.

A practical framework for personal AI use#

Over the past two and a half years, I've developed a simple set of design principles and questions that guide how I choose and use AI tools.

1. Strengthen your capability, don't replace it#

Use AI to support how you think, not to outsource it completely.

As a writer, I found that generating content with AI weakened my voice. Instead, I now use it to structure ideas, format documents, and review and refine.

Principle: Keep authorship and thinking human-led.

2. Be specific about what you used#

Name the tool and the task. "I used Claude to summarise a 20-page report into three bullet points for my executive update" is more useful than "AI helped me with a report."

Principle: Specificity builds understanding and trust.

3. Check what you're handing over#

Before connecting any tool to your email, calendar, or documents, ask three questions. What data does this need. Where does it go. Who else can see it.

Principle: Data literacy is part of AI literacy.

4. Have a plan before you pick a tool#

Start with the problem, not the solution. What are you trying to achieve. What do you already know. What gap are you trying to close.

Principle: A tool is only as good as the clarity of the task you give it.

5. Look for the real price#

The bright yellow "special" tag isn't always a special. Free tiers often come with data harvesting. Cheap subscriptions may lack the security or support you need for professional work.

Principle: The real cost includes your data, your time, and your reputation.

What this means for your work#

These principles are not about avoiding AI. They're about using it well.

The people who will thrive in the Intelligence Age are not those who use the most tools. They are those who use the right tools, for the right tasks, with clear intention.

Start with one tool. Learn it properly. Be specific about what it does for you. Check what you are handing over. And always keep your own thinking in the driver's seat.

---

Meg Smith is CEO and AI Strategist at Cloverbase. To discuss this article or work with me, contact me at Cloverbase.

---

Share this article#

https://m1.nz/gz9qe2s