Smart Event Studio

Someone DMed me last week after I shared a tool I'd built in Replit. "It's really cool that you're able to use AI to solve that problem," she wrote. "These are the type of innovations our industry desperately needs."

It was cool to get the comment, but I also remember thinking, "this explains so much about our industry."

What I built wasn't innovative. It was a URL validator that I use to check exhibitor URLs to repair malformed URLs, find dead links, and — most importantly — keep me from clicking on malware and infecting my computer. Also, manually checking thousands URLs for large trade shows is not exactly the kind of work I love doing. That's the whole story: problem, annoyance threshold, tool.

"The way you do anything is the way you do everything." My business coach said that to me recently about something else entirely. But I think it fits in here, too.

Name the problem before you touch a tool

I see a lot of people approaching AI backwards. They wonder what AI can do, then go looking for problems to match it to. Which is exactly how you end up with a prompt that generates your weekly status update and not much else.

Try this instead: pay attention to where your time goes. Not broadly — specifically. Not "data management is a mess" but "every event before we launch I have to spend 40 hours manually verifying 4000 entries and I lose my will to live." It's THAT level of specificity is that makes the difference between a tool that genuinely helps and a what is nothing more than productive procrastination.

My Replit validator came from that kind of "problem → annoyance → tool" scenario. So did a test plan generator I put together recently, a PDF-to-Markdown converter, a Calendly sync tool, and a requirements\ validator. None of them are sophisticated. All of them came from asking: what is the specific thing that keeps wasting my time, and what would it look like if it just... didn't?

ChatGPT is the entry point, not the destination

Generative AI (ChatGPT, Claude, Gemini, Grok, etc) has come a really long way really fast — particularly in the past few months.

Most people get as far as chat and they stop. But it's easier and easier for non-technical people to build true tools that will save time and energy — and they have different data privacy implications.

The first is the chat interface layer — ChatGPT, Claude, Gemini. Fast, accessible, good for thinking out loud, drafting, summarizing. The tradeoff is that data handling varies by platform and plan, so know your terms before client data goes anywhere near it. And regardless of platform: no PII, no confidential client information, full stop.

The second is the structured environment layer — Claude Projects, Custom GPTs. These give the AI persistent context: your templates, your voice, your way of working. Less setup than building something from scratch, more consistency than a blank chat window every time.

The third is build-your-own — Replit, Lovable, APIs. You describe what you need in plain language, and the tool builds it. APIs often have stronger data controls than consumer interfaces, which matters if you're doing anything client-facing. This is where I've been spending time lately, and I want to mention it specifically because people assume it requires a development background. It doesn't. It requires knowing what you want.

You don't need to understand it to start

My business coach's observation — the way you do anything is the way you do everything — applies here in a specific way. The people making real progress with AI already had a habit of noticing what wasn't working and looking for a better way. AI didn't give them the habit — it just made the feedback loop faster.

You don't need to understand large language models. You don't need to learn to code. You need to know\ what's actually taking up your time — specifically — and be willing to try something before you're certain it'll work.

The URL validator took an afternoon. It saves me hours, my sanity, and my computer every time I use it. That's the whole return on investment calculation. Start there.