Your team is probably already using Claude. Someone’s drafting emails with it, someone else is summarising long documents, and a few people quietly lean on it more than they’d admit. That’s the normal pattern now.
The real gap isn’t whether your staff use AI. It’s the difference between using it occasionally and getting reliable, safe value from it. Claude, made by Anthropic, is one of the main AI assistants businesses reach for alongside Microsoft Copilot and ChatGPT. If your team has moved past the novelty stage, here are four tips to get more out of it without creating new risks.
1. Set up a Project so you stop re-explaining your business
The biggest time waster with any AI tool is starting cold every time. Claude’s Projects feature fixes that. A Project is a workspace where you load the documents, context and instructions Claude should always keep in mind, so every chat inside it already knows your background.
Set one up for a recurring job, say client proposals or monthly reports. Add your templates, your tone of voice, a couple of past examples, and custom instructions that explain how your business works. On the paid plans each Project holds a large amount of reference material, so you’re not pasting the same briefing in over and over. Your team can share their best chats too, so a good prompt spreads instead of living on one person’s screen.
2. Tell it who it is and what good looks like
Vague questions get vague answers. The quickest improvement most people can make is to give Claude a role, some context, and an example of what a good result looks like.
Instead of “write a follow-up email”, try “You’re writing for a New Zealand IT company. Write a short, friendly follow-up to a client who asked about pricing last week and hasn’t replied. Keep it under 120 words and finish with a clear next step.” The more you tell it about the situation and the format you want, the less you’ll edit afterwards. And if the first answer misses, say what to change rather than starting over. It remembers the thread.
3. Keep client data out of consumer AI
This is the one that matters most for a business. Free and personal AI accounts can use what you type to help train future models. That’s fine for a recipe. It’s not fine for client records, financial details, passwords, or anything covered by the Privacy Act.
The fix is a business-grade plan. Claude’s Team and Enterprise plans don’t train on your conversations by default, and they add admin controls and single sign-on so you can see who’s using it and keep accounts tidy. If staff are already pasting work information into personal AI logins, that’s worth sorting sooner rather than later. It’s the same thinking we bring to IT consulting: get the setup right so the convenient option is also the safe one.
4. Treat it as a fast first draft, not the final word
Claude is quick and often very good, but it can be confidently wrong. It will state a wrong fact or invent a detail in the same calm tone it uses for correct ones. So use it to reach a strong first draft fast, then apply your own judgement.
For anything that goes to a client, gets published, or informs a decision, check the facts and read it back as yourself before it leaves the building. Think of it as a capable junior who works at speed and needs a quick review, not an expert whose word you take on trust.
What this means for your business
Used well, Claude and tools like it take real time off everyday work: writing, summarising, research, and first drafts of documents. Used carelessly, they create privacy and accuracy risks that land back on the business owner.
The difference is mostly setup and habits. Picking the right plan, switching off model training, building Projects for the jobs you do often, and agreeing a few simple rules about what staff should and shouldn’t put into AI. None of that is difficult, but most small teams haven’t done it yet.
How Project IT can help
We’re now helping clients put exactly this in place as an AI advisory service. That means advising on which AI tools fit your business, setting them up on the right plans with proper security and admin controls, building Projects and prompts around your real work, and giving your team simple guardrails so they use AI safely. It sits alongside our IT consulting, just pointed at AI rather than servers and networks.
If you want your team using AI properly, without the privacy headaches, Book an IT Review and we’ll map out a sensible starting point.
Quick questions
Is it safe to use Claude for work?
On a business-grade plan, yes. Team and Enterprise plans don’t use your conversations to train models and add admin controls. The risk comes from putting client or sensitive data into free personal accounts.
Claude or Microsoft Copilot, which is better?
They overlap and plenty of businesses use both. Copilot is built into Microsoft 365, while Claude is strong at writing and working through long documents. The right mix depends on your tools and the work you do, which is part of what we help sort out.
Do we have to pay for it?
There are free tiers, but for business use a paid Team plan is worth it for the privacy protection, admin controls and higher limits.