When I decided to automate parts of MTL Hiking, I didn’t know where to start. I didn’t know what to automate. I didn’t even know what was possible.
So I did research. For days. I’d sit down to start, get overwhelmed, give up, and end up doing nothing.
What I tried
Every article and video pointed to another tool. ChatGPT or Claude. Then inside Claude, Opus or Sonnet. Then agents, platforms, and a new version of everything every few weeks. It felt endless.
I kept trying to pick the best option before doing anything. Every comparison opened three more questions. I saw automation as one huge mountain I had to climb all at once, and I couldn’t find the first step.
The one task I picked
My hikes depend on carpool drivers. I have more than 70 in my pool.
Their sign-ups lived in a Google Sheet. I shared a list of upcoming hikes and drivers wrote their name next to the ones they wanted. The week before each hike, I built its page manually in Notion and sent it to our drivers group so they could sign up on Eventbrite. Then my admin and I spent the week chasing the ones who hadn’t.
So I stopped comparing. I picked Notion because I was already using it. A new tool would only have made things more complicated.
How Claude helped
I told Claude how I handle driver sign-ups today and what I needed instead. It suggested a solution, and we built it together in 3 to 4 hours.
- One form. Full name, email and the hikes. A driver can pick several at once and see each one’s time and difficulty first.
- One database. Every answer lands in a Driver Sign-ups database in Notion.
- One calendar. A public calendar shows drivers what’s coming.
- One email. When a driver signs up, they get their Eventbrite link so they can claim their spot.
Even if I end up changing it later, at least I did one thing.
What failed
The real failure came before any of this. Days of research that ended in nothing. I could have built one small thing on day one.
And the form isn’t proven yet. Zero drivers have used it so far. The email is built, but I haven’t tested it with a real driver. That’s the part I care about most, because it’s what should stop my admin and me from chasing people to sign up.
Anything a driver touches also has to explain itself. Not every driver reads instructions, so a form that makes sense to me can still confuse them. Testing will tell.
What I do now
That form was a victory, even before its first sign-up. It’s what got me automating more of the business. Here’s how I start anything new now.
- One task. Pick the smallest thing you do by hand every week. Mine was a sign-up sheet.
- The tool you already use. You can switch later, and you’ll know much better what you need by then.
- Tell Claude what you do today. Then what you need. Let it suggest the solution.
- Working before better. Get the simple version running. Then add one piece at a time.
