A friend called me last week with a problem I see all the time.

He runs a subscription business with over 1,000 customers and was moving from PayPal to Stripe.

The challenge?

He had to manually contact customers when their subscriptions were expiring so they could convert to the new system.

His best conversions came through phone calls and texts, not email.

He still works a full-time job, so he spends hours every morning just figuring out who to contact.

Here's what was killing him: customer data scattered across three different systems, and no easy way to see the complete picture.

I've seen this exact problem dozens of times as a management consultant.

Leaders are spending valuable time wrestling with spreadsheets instead of growing their business.

But this time, I walked him through something different.

Why Most People Fail Before They Start

Most people make their biggest mistake before they ever use AI.

They jump straight into ChatGPT or Claude without thinking through what they actually want to accomplish. Then it becomes this endless back-and-forth without any structure.

From my experience, you need to have a solid foundation before you decide to use AI.

Here's the process I walked him through:

Step 1: I Had Him Map His Data Sources

Before touching any AI tool, we identified what he was working with:

Each system had pieces of the puzzle, but none gave him what he actually needed to take action.

Step 2: We Defined the Real Goal

For every PayPal customer, he needed to know:

Step 3: I Used AI to Analyze, Not Execute

Here's where my approach differs from that of most people.

I put all his data into Claude Opus. But instead of asking it to solve the problem, I had it analyze what we were working with and show me the gaps.

It created a comprehensive breakdown showing:

That analysis gave me an idea for something much better than manual tracking.

Step 4: We Built an Automated Solution

This is where my management consulting background proved useful.

Instead of having him manually update spreadsheets every day, I designed an automated system. I had him create a Stripe API key, then I used AI to help me build a Google Apps Script that:

The Results

Now every morning at 7 AM, he gets an automated email with exactly what he needs:

What used to take him 2-3 hours of manual work now happens automatically.

And here's the key: there's no AI involved in the daily operation. We used AI to build the solution, not to be the solution.

How I Apply This to Other Business Challenges

From my consulting work, I know that this same approach works for any situation where you manually combine data from different sources.

The framework I use:

Why This Matters for Your Business

You don't need AI to do your work.

You need AI to help you eliminate work that shouldn't exist in the first place.

My friend's next step? We're adding automated email campaigns so the system doesn't just identify who to contact, it reaches out automatically with personalized renewal messages.

Ready to Stop Wrestling with Spreadsheets?

Try this strategic approach with your data challenge.

Start by mapping what you have and what you need, then use AI to help you design a system that gives you answers instead of more work.

From my experience, this is how you actually save time, rather than just creating more sophisticated busywork.

Marvin

PS: Three ways THINK can help you slash manual work: