AI is no longer this far-off, sci-fi buzzword. It’s here. It’s real. And it’s already running quietly in the background of more businesses than you might think.
In fact, 78% of companies are using AI in some shape or form right now.
That number was 55% just last year.
And by the end of 2025, we’re expected to hit a cool $644 billion in generative AI spend globally. That’s not hype. That’s momentum.
But if you’re not Google or Meta with a warehouse full of engineers and GPUs, where do you even start?
Good news: you don’t need to be a tech giant to get real value out of AI.
You can ease into it with a super simple framework I use all the time when consulting:
Level 1: Use AI tools
Level 2: Automate with agents
Level 3: Build your own systems
Let’s break it down.
Level 1: Use AI Tools for Quick Wins
This is the toe-in-the-water phase. You’re not building anything fancy—just using tools that already exist. Think ChatGPT, Claude, Perplexity, Grok. No coding. Just vibes.
You give it data, and it gives you answers.
- 📝 Got a marketing campaign to write? Load in your sales data and ask it to draft an email to your top-performing demographic.
- 📣 Need social content? Ask it to write 10 Instagram captions based on your Q2 report.
- 🧾 Drowning in receipts? Paste them in and say, “Sort these by category and flag any expenses over R5k.”
It’s like having an extra brain that doesn’t sleep or get bored.
And the best part? Tools like this are either free or ridiculously cheap. Set up takes minutes. You can start using them today.
Downsides?
These tools aren’t great for teams. You have to upload fresh info every time, and there’s no live connection to your systems. It’s manual. But that’s okay—this level is about learning and spotting where AI can save you time.
Level 2: Agent-Based Automation (hello n8n)
Once you see the value, the next step is automation.
This is where tools like n8n come in (open-source, super flexible). Now you’re not just using AI—you’re wiring it into your workflows.
Imagine this:
- You’re running an e-commerce store.
- An n8n agent pulls your Shopify sales data hourly.
- It runs that through ChatGPT to spot trends.
- Then it updates a Google Sheet and pings your team on Slack if something interesting pops up.
That’s Level 2.
Or let’s say you’re a content team.
An agent can scan RSS feeds for industry news, look at your Trello content calendar, and then generate article ideas. You wake up to 3 draft blog posts already written and waiting.
Or your finance team connects Xero to a dashboard that auto-summarizes last week’s spending. No more spreadsheets.
What’s the catch?
Set-up takes time. You need to think in workflows, and you’ll spend a bit of time debugging and tweaking. Also, if your tools change (API keys, login methods, etc.), your agents might break.
But it’s worth it. The amount of repetitive tasks we all do that we could automate would save heaps of time.
Level 3: Build Custom Low-Code Systems
Now we’re getting serious.
This level is about building your own tools. Not from scratch—just enough to solve real problems your way.
This is where you can take tools like Cursor and actually build a tool for your business.
One tool I've thought about is something as simple as an expense claim app.
An interface where employees can upload receipts, AI sorts them and then spits out a report to accounts to process.
So rather than having a whole human processing each slip, AI does the repetitive stuff, and someone in accounts can just check it and process it.
Imagine that person could spend the rest of their time doing something far more meaningful for their business.
Think about all the other little tasks in your business that you pay people to do. How much of that could just be automated so that your employees could do work that actually drives the business forward.
The upside?
It’s built for you. It scales with your business. And it can reduce errors by up to 90% and costs by 30–50%.
The downside?
There’s a learning curve. And you’ll want to think about data privacy/security. But if you’re running a growing team or juggling multiple systems, this is the future.
So, Where Do You Start?
Start small.
Test ideas.
Use what’s already out there.
Then automate what works.
Then build what’s missing.
Level 1 → Level 2 → Level 3
That’s the roadmap.
We’re still early—but not that early.
The businesses that move now are going to outpace the ones that wait.
AI doesn’t replace humans—it amplifies them.
And the ones who learn how to work with it are going to win.
Catch you next week,
—Shawn