A Note from Andy Larin, CEO
Artificial intelligence isn't coming to your business — it's already there. A recent global study from KPMG and the University of Melbourne found that 58% of employees are already using AI tools at work, with a third doing so weekly or daily. That number is only going up.
What isn't keeping pace is governance. The same study found that only 34% of organizations have a policy guiding how AI should be used. That means most businesses are running an experiment they didn't sign up for, with tools they may not have approved, and data they may not realize is at risk.
This month's Cyber Insider is about changing that. We'll cover how to get real, practical value out of AI tools — and give you a framework for your own AI policy to help protect your business, your clients, and your reputation.
AI done right is a genuine competitive advantage. Let's make sure you're set up to benefit from it.
FEATURED ARTICLE
From AI Zero to AI Hero: How to Build an AI Prompt That Actually Works
You just wrapped up your quarterly business review. You've got rough notes from two hours of discussion, and you need to turn them into a board update — fast. You open your AI tool and type:
"Summarize my meeting notes in an email."
And you get back... something. Technically it's an email. But it's generic, it's missing the point, and you're going to spend the next 20 minutes rewriting it anyway.
The problem? It’s an incomplete prompt getting an incomplete result. Think of AI like a genie — incredibly powerful and eager to help, but if your “wish” is vague or open to interpretation you may not get what you were expecting.
Here are four simple additions to take your prompt from zero to hero:
- Task: Be specific about what you actually want.
"Summarize these meeting notes in an email to be read by our board members."
Naming the audience changes everything. The AI now knows to write for executives, not a general audience. Good start — but we can do better. - Context: Tell it what matters.
"The board is particularly interested in revenue performance against targets and any operational challenges heading into next quarter.”
Now the AI isn't guessing at what's important. You've pointed it at the right information. The summary will be sharper and far more relevant. - Format: Don't leave structure to chance.
"Open with a brief summary paragraph, followed by three clearly labelled bullet sections: Wins, Goals for Next Quarter, and Lessons Learned."
Without this, you get whatever structure the AI feels like using. With it, you get exactly the layout you need — ready to drop into an email with minimal editing. - Tone: Say what you need, including what you don't want left out.
"Keep the tone professional and avoid technical jargon. Do not minimize areas for improvement — the board expects candour."
That last sentence matters more than it might seem. AI has a natural tendency to prioritize a non-confrontational tone over harsh accuracy. If you need an honest assessment — and in a board update, you usually do — you have to ask for it explicitly.
Here's your completed “AI hero” prompt:
"Summarize these meeting notes in an email to be read by our board members. The board is particularly interested in revenue performance against targets and any operational challenges heading into next quarter. Open with a brief summary paragraph, followed by three clearly labelled bullet sections: Wins, Goals for Next Quarter, and Lessons Learned. Keep the tone professional and avoid technical jargon. Do not minimize areas for improvement — the board expects candour."
Same meeting notes. Same AI. Completely different result.
This pattern works for almost anything — lead with a clear task, add context, define the format, and be explicit about tone. You'll spend far less time fixing AI output, and a lot more time benefiting from it.
EYE ON A.I.
Your Employees Are Already Using AI at Work. Here's What a New Study Found.
A sweeping new study from KPMG and the University of Melbourne drew on insights from 48,000 respondents across 47 countries — and the results should be on every business owner's radar.
- 58% of employees are already using AI at work — a third of them weekly or daily
- 70% are using free, public tools rather than employer-approved ones
- 44% admit to using AI in ways that violate their organization's policies
- 66% of organizations have no policy in place to govern any of this
The concern isn't that employees are using AI — it's how. Most are using whatever tool is free and familiar, with little awareness of what happens to the data they put into it.
The Bottom Line: Your employees aren't waiting for a company policy before using AI. Don't wait until something goes wrong before you create one.
BULLETPROOF YOUR BUSINESS
Your Business Needs an AI Policy. Here's What to Put In It.
As this month's Eye on A.I. highlighted, 66% of organizations have no policy governing how employees use AI. If you're in that majority, it's worth fixing sooner rather than later. The good news? You don't need a 30-page legal document. You need a clear, practical policy your team will actually read, understand, and follow. Here's what it should cover:
- Your Organization's Official Position on AI
Start with the big picture:
Is AI use permitted at your organization?
Under what conditions?
Who does the policy apply to?
This doesn't have to be complicated — even a brief statement of intent sets the tone and signals to employees that AI use is something your business takes seriously. - Approved Use Cases — and Prohibited Ones
Be specific about what AI can and can't be used for in your business.
Approved uses might include:
drafting communications
summarizing documents
brainstorming marketing ideas
Forbidden uses may include:
creating fake or defamatory content
making HR decisions
infringing on third-party intellectual property rights - An Approved Tools List
To help eliminate shadow AI – name the tools and accounts your team is permitted to use — and include a simple process for requesting approval of new ones. - A Clear Data Boundary
Even with approved tools, define which data types are still off-limits unless explicitly authorized. This list might include:
client personal information
employee records
financial data
credentials (passwords, login information)
anything covered by your confidentiality agreements
The more explicit this list, the less room there is for well-intentioned mistakes. - Ownership and Consequences
Give your policy some teeth.
Name the person or role accountable for AI governance in your organization
Outline what happens when the policy is violated.
Accountability is important. A policy without consequences is just a suggestion.
Every organization is different — your industry, size, client obligations, and regulatory environment will all shape what your policy needs to address. The five points above are a starting framework, not necessarily a complete solution. Remember to build in a scheduled review — annually at minimum — as AI usage will continue to grow and change.
→ Not sure where to start? allCare IT can help you build or review an AI Acceptable Use Policy that fits your business. Contact us to get started.
TECH TIP
Check Your Facts: Your AI Tool Might Be Giving You Obsolete Information
Remember buying encyclopedias? The information was accurate and thorough — right up until the moment they went to print. Population figures, currency values, even national boundaries that were correct on publication day became quietly misleading — or just wrong — a few months or years later.
AI tools are similar. They are trained on data up to a specific point in time — called a knowledge cutoff — and anything that happened after that date simply isn't in there. This means your inquiry might return a confident, detailed — and completely outdated — answer.
What to do: For anything time-sensitive or prone to change, always verify AI output against a current source. When in doubt, ask your AI tool directly: "What is your knowledge cutoff date?" Many tools will tell you. If not, check documentation on the specific model you’re using. Your subsequent fact-checking might take a bit more time, but you’ll be confident that you’re working with the most current information available.
SERVICE SPOTLIGHT: HATZ SECURE AI
There's a Better Way for Your Team to Use AI — And Your Data Stays Out of It.
This issue has covered a lot of ground — how to get better results from AI, why your team is probably already using it without a policy, and what that policy should include. But there's one practical question underneath all of it: if free public AI tools are the problem, what's the alternative?
Meet Hatz Secure AI — an enterprise-grade AI platform allCare IT deploys specifically for businesses like yours.
Here's what makes it different from ChatGPT or whatever your team is using right now:
- Your data never trains the AI models — it's written into the agreement
- SOC 2 Type II certified — independently verified security controls
- 58+ AI models in one platform — no more managing multiple subscriptions
- Full audit trails and access controls — you see exactly who's using AI and how
- Built-in learning module — your team goes from beginner to confident user
Everything we've covered this issue — approved tools, data boundaries, governance — Hatz is built to deliver all of it out of the box.
→ Ready to see it in action? Book a discovery call with allCare IT and we'll show you exactly how Hatz works for your business.