Working with AI

You don't need an AI strategy. You need 30 days.

Most businesses overthink AI. Here's what I'd do in the first month instead.

By Sarah Wood, Creative Sauce AI

One of the most common things I hear is: "We know we should be doing something with AI. We just don't know where to start."

It's understandable. Every day there's a new model, another AI tool and another LinkedIn post telling you that if you're not using agents, automation or autonomous workflows, you're already behind.

You're not. In fact, I'd argue the opposite.

Most businesses don't need an AI strategy. They need thirty days.

Thirty days to understand where AI genuinely helps, where it doesn't, and where it can start saving time without creating new problems. Here's how I'd spend them.

Week 1: find the repetition

Don't start by choosing a tool. Start by noticing where your team keeps repeating the same work:

Customer emails
Meeting notes
Writing proposals
Review replies
Research
Social media

If something happens several times a week and follows a similar pattern, write it down. You're not looking for the biggest opportunity. You're looking for the easiest win.

Week 2: use one general AI assistant

Pick one tool. ChatGPT. Claude. Gemini. It genuinely doesn't matter very much at this stage.

Spend a week using it for everyday work. Rewrite an email. Summarise a document. Brainstorm ideas. Draft a proposal.

The goal isn't to become an AI expert. It's to understand where it naturally fits into your working day. Don't download six different platforms. Learn one properly.

The geeky bit

People often spend hours comparing models before they've solved a single business problem. In reality, the model is rarely the biggest factor. Your instructions, the documents you give it, and the workflow around the AI usually make far more difference than whether you're on one model or another. A clear brief and the right context turn an average model into a useful colleague; a vague prompt makes the best model in the world look mediocre. That's why businesses often see bigger improvements by refining how they use AI than by constantly switching tools. The skill you're building this month isn't picking the winner. It's learning to ask well and hand over the right information.

Week 3: measure one thing

Choose one task. Time yourself doing it normally. Then do it again with AI. If the quality is good enough and you've saved twenty minutes, you've found something worth keeping.

Don't measure prompts. Don't measure tokens. Measure time saved. That's the number that matters.

Week 4: decide what deserves more investment

After a month you'll know far more than you do today. Some tasks won't suit AI. Some will save a few minutes. One or two will probably make you wonder why you've been doing them manually for years.

Those are the ones worth building properly. Maybe that's an automation. Maybe it's a dashboard. Maybe it's a business assistant connected to your systems. The important thing is that you've earned that decision through experience, not guesswork.

What I'd avoid

Don't buy five subscriptions.
Don't build an AI agent because everyone else is.
Don't automate a process you don't understand.
Don't expect AI to fix disorganised data.
And don't worry about doing everything at once.

The businesses getting the most value from AI didn't transform overnight. They found one good use case. Then another. Then another. Thirty days later, they weren't talking about AI anymore. They were simply getting more work done.

Where to start

If you're wondering where to start, don't start with the technology. Start with the work you're already doing. Everything else becomes much easier after that.

Would rather it just happened? The Growth Toolkit is six tools that grow your business quietly in the background — from £13 a month for all six, part of the Creative Sauce Growth Toolkit. See how it works →

If you'd like a hand working out where AI genuinely fits, or want to skip straight to building the one job worth building, that's exactly where we come in.

Book a quick chat →

Common questions

Do I need an AI strategy before I start?

No. Most businesses don't need an AI strategy, they need thirty days. A month of using AI on real work tells you where it genuinely helps far better than any strategy written before you've tried it. Find the repetition, learn one assistant properly, measure the time saved, then decide what deserves building.

Which AI tool should I start with?

It barely matters at the start. Pick one general assistant, ChatGPT, Claude or Gemini, and use it properly for a week on everyday work. Learning one tool well beats downloading six. The model is rarely the biggest factor, your instructions and the workflow around it make more difference.

How do I know if AI is actually helping?

Measure one thing: time. Take one task, time yourself doing it normally, then do it again with AI. If the quality is good enough and you've saved twenty minutes, you've found something worth keeping. Don't measure prompts or tokens. Measure time saved.