FridAI Pomelli and Lyria

In this episode of Business Brain, we dig into FridAI territory with updates, experiments, and smarter AI workflows. We kick things off with a Press Row Hoops app update, then pivot into something far more powerful: upgrading our custom AI instructions. Instead of settling for generic outputs, we refine how ChatGPT behaves—focusing on decision-making style, pushback tolerance, clarity versus depth, and minimizing fluff. We share a tight prompt that forces better performance from ChatGPT 5.2, helping us get sharper answers with fewer wasted cycles. The goal? Make AI think more like we do.

We also explore Google’s Pomelli and what it signals about where AI tools are heading. The bigger takeaway is this: when we intentionally design how AI supports us, we reclaim leverage and reduce friction. We stop accepting default settings and start engineering better results. That’s how we build businesses—and lives—on our terms. That’s the path to a Charmed Life.

• 00:00:00 Business Brain – The Entrepreneurs’ Podcast #731 for Casual FridAI, February 27th, 2026
  • February 27th: Anosmia Awareness Day
• 00:01:22 David-Press Row Hoops app update
• 00:05:27 Ian-Update your custom AI prompt/instructions, and here’s a prompt:
  • Review my existing custom instructions.
    • Identify what works well and what is outdated, redundant, unclear, or ineffective for ChatGPT 5.2.
    • Ask only the minimum number of high-value questions needed to meaningfully improve the instructions.
    • Questions should focus on decision-making style, pushback tolerance, clarity vs depth, and how I prefer ChatGPT to behave when uncertain.
    • Do not ask generic preference questions or enter question loops.
  • After I answer, draft a replacement custom instruction that reflects my preferences and how I actually use ChatGPT.
  • Requirements for the new custom instruction:
    • Maximum 1500 characters, including spaces and line breaks.
    • Written as direct instructions to ChatGPT, ready to paste.
    • Optimized for ChatGPT 5.2 behavior.
    • Clear, human-sounding, and practical.
    • Favor correctness over politeness.
    • Avoid fluff, buzzwords, and assistant-sounding language.
    • Minimize unnecessary clarifying questions.
    • Default to direct answers with light supporting rationale.
    • Be opinionated when useful and push back on weak assumptions.
    • Reuse prior context and preferences automatically unless I say otherwise.
  • Constraints:
    • Do not draft the final instructions until after I answer your questions.
    • Do not explain your reasoning unless I ask.
    • Deliver one final custom instruction, not multiple options.
• Sponsors
  • 00:08:09 SPONSOR: Granola is an AI-powered notepad built for the way real people actually meet, and it integrates seamlessly into the video conferencing tools you already use. Try Granola totally free for three months – just head to granola.ai/brain.
  • 00:10:03 SPONSOR: Fundera from NerdWallet – A free, easy-to-use platform that lets you compare real financing offers from trusted lenders — all in one place. Visit NerdWallet.com/BRAIN to learn more and talk to a real person!
• 00:11:26 Google Pomelli
• 00:18:00 Business Brain 731 Outtro <https://businessbrain.show/>
  • Tell Your Friends!
  • Review Business Brain
  • Subscribe to the show
  • feedback@businessbrain.show
  • Call/Text: (567) 274-6977
  • X/Twitter: @ShannonJean & @DaveHamilton, & @BizBrainShow
  • LinkedIn: Shannon Jean, Dave Hamilton, & Business Brain
  • Facebook: Dave Hamilton, Shannon Jean, & Business Brain

The post FridAI Pomelli and Lyria – Business Brain 731 appeared first on Business Brain - The Entrepreneurs' Podcast.