My Claude is on steroids now.

Big thanks to Mitchell Jones and his company, Lava. (Mitchell has been a client of Wonderfish, and he has been doing amazingly well on podcasts.

I spent last night in NYC at one of Lava's live build sessions. I walked in with a laptop and a problem, and walked out with a working AI agent.

Here was the problem:

When my clients appear on podcasts, those episodes go live on Spotify, Apple Podcasts, and YouTube. But the media monitoring tools we use don't always catch them on time.

We try to stay on top of it by emailing podcast hosts after recording, but they don't always tell us when the episode actually goes live.

So I built a simple solution: a daily podcast tracker.

Here's what it does:

→ Every day, it searches across multiple platforms to find new podcast episodes → It focuses only on the last 24 hours and removes duplicates → It separates real episodes from things like social posts or promotions → It organizes everything into a clean report with the episode name, show, date, and link → It sends that report directly into Slack and also gives a link to a Lava page with the report

What surprised me most:

The hardest part of building something like this isn't the AI logic. That part is actually pretty simple.

The hard part is all the behind-the-scenes setup—managing logins, connecting tools, and dealing with limits across different platforms.

Lava simplifies all of that into one connection. Instead of wiring everything together, you can just focus on solving the problem.

And this isn't just for podcasts. You could use the same setup to track things like earnings reports, competitor pricing, regulatory updates, or customer issues. Same structure, different use case.

I plan to build a few other tools that do daily tasks tied to my HubSpot workflow.

Huge thanks to the Lava team for being incredibly hands-on throughout the build.

If you get the chance to attend one of their sessions, bring a real problem. You'll likely walk out with a working solution.