This week's Hotseat opened with a very heartfelt (albeit slightly laggy) rendition of happy birthday in celebration of my special day! Exactly the kind of warmth I’ve come to admire about this community. It was a nice full room, with many familiar faces, one new one, and the kind of energy that signals when people feel genuinely happy to be back in a shared space.

The theme underpinning almost everything this week was AI: how best to use it, how to protect your IP from it, and how to stay grounded while the tools continue to shift and change almost daily. Beneath the tech talk was the usual thread of community, real progress shared honestly, and people helping one another soundboard and problem solve.

Package what's already in your head.

One idea I keep coming back to from the session: Most of us are sitting on far more intellectual property than we realise! Nik put it well when he said that even ten years in a job leaves you with a wealth of expertise worth packaging.

The practical version of this is a simple digital product, a PDF or template, sold for personal use and offered again with a white-label license at a higher price. It turns knowledge you already have into something with double earning potential.

Protect the work only you can offer.

Margaret was approached by an long time listener and AI bot developer interested in utilizing her podcast content for a niche chatbot. This sparked useful dialogue around questions more of us may face in the future.

Here’s what came up:

●Your most valuable material is actually the work you don't publish. So give snippets, but keep the depth of your knowledge for your community and your one-on-one work.

●If you go the legal licensing route, this can be a real income stream. If your content is used to run a bot, that's an ongoing fee for as long as it's live, not a once-off payment.

●Consent and clear terms matter in cases like this, even if other LLMs scrape broadly without consent. Doing it properly in instances where you have control is what separates a partner from a problem.

Foundations first, automation second.

Pearl admitted that she’d started building an agentic AI to run some daily tasks, then lost interest when it got overly complicated. Sound familiar?

Simone's advice offered a grounding take here:

●Decide what you actually want it to do before you build it. Work backwards from the goal.

●Get your task management working as a simple system first, then let AI sit on top of it.

●A basic version that runs, beats an ambitious one that stalls. Iterate from the basic version.

Master a few tools instead of chasing all of them.

When I walked the group through how I use Claude, from usage credits and connectors to reusable design systems and scheduled actions, the risk was making it feel like everyone needs to learn everything. Nik reframed it perfectly.

The AI landscape is constantly shifting. So the move isn't to chase every new release, it's to rather work with tools you already trust and let them grow with you. That said, here are at least two useful tips that will save you tokens along the way:

●Draft messy thinking in a token free tool, have it write you a tight prompt for the tool that requires token use. Shorter, clearer prompts cost fewer tokens and hold the thread better.

●Paste text instead of screenshots. Images cost far more tokens to read than plain text.

What we're each doing in the next week or two

Refer 3x friends who join and you get a lifetime FREE membership. Send them this email as a sneak peak!