Be Fickle & Hard to Please—AI as a Creative and Strategic Partner
Part 4 of our 20 part series
Welcome to AI as a Creative and Strategic Partner—Funmentum™ Style!🎉🎉🎉
Our new series on harnessing Generative AI like Chat GPT, Copilot, Gemini, etc. for both fun and breakthrough results. Many people use GenAI for low level tasks like summarizing meeting notes, writing simple emails, researching industry trends, etc.
This series is focused on the opposite—Gen AI is strongest as a high level thought partner. Not replacing you or doing your work in an average/generic way, rather pushing your thinking as a strategic challenger, creative muse, and feedback focuser. All to create higher quality work, which is approved faster, with less and time/effort. And it’s a lot of fun!
Make sure to subscribe and follow along—new posts every Friday at 5am PT/8am ET for the next 16 weeks!
Chapter 1 of 4: Mindset is the Magic
For the first 5 posts, we’ll focus on the right mindset for working with AI. Namely, way more context and back and forth iteration than you expect.
#4: Be Fickle & Hard to Please
In our first three posts, we covered treating AI like a diary, uploading everything you've got, and giving the AI the perfect role to play. Now we're tackling something that might feel a little unnatural: being hard to please and pushing for more.
New 2024 research from
and out of Stanford (download here) shows that teams that used ChatGPT to ideate on work challenges came away with more average ideas than similar groups using traditional, non-AI ideation sessions (aka what I host all the time). BINGO—AI skeptics, they proved what we all intuitively knew! But wait…👀🤫The Secret to Working with AI
Teams that got average ideas with AI treated it more like “an oracle”🔮, as the study authors call it. Those teams saw ChatGPT’s first answer, thought it seemed pretty good, and said, “OK, great. Now we can go to lunch”. However, the teams that were dissatisfied with ChatGPT’s output, and pushed it to keep iterating ideas, outperformed all teams by a mile, including the non-AI, traditional ideation session teams.
❌Bad Example - The Satisfied Customer👌
Gets AI response ➡️ “This seems fine.” ➡️ Uses mediocre ideas ➡️ Wonders why AI isn’t that helpful😩
✅Good Example - The Demanding Director🎬
Gets AI response ➡️ 1st iteration: “Ah, I forgot to tell you ______, can you redo it?” ➡️ 2nd iteration: “None of this is feeling very creative or interesting, can you give me 10 more that I haven’t seen?” ➡️ 3rd iteration: “Getting closer with idea #4, give me more like that.” 🔁 Keep going until you hit gold💡
Here's a personal example: last week I was writing a workshop outline, but the exercises it suggested were all ones I’d seen before, not that interesting. A couple of years ago I would have said, “See—AI isn’t actually that useful,”🙄 and just gone back to my old way of doing it.
But now, instead of settling, I just kept pushing, asking it to give me more options over and over. Then I'd quickly voice-record my thoughts, reviewing the ideas out loud in real time. Not carefully crafted feedback, just rapid, “in the moment” reactions. After about 4 rounds of this, I got to the ✨perfect exercise✨ that made it into the final workshop. The key was speed—this whole process took me maybe 10 minutes. I didn’t get bogged down in analyzing every mediocre output from the AI, I just pulled out the interesting ones, and kept pushing, rapidly, until I struck gold.
The most important thing is to just flow with it. Many people go wrong working with the GenAI as just a “single perfect prompt”, almost like dusting off a video game cartridge and uploading it to the machine. It’s much more conversational, exploratory, finding rabbit holes and pushing.
🏅Pro Tips for Being Productively Picky:
🌊 Ask for quantity - Keep the iterations flowing. "Give me 10 more" is your new best friend
🚀 React fast - Don't spend 10 minutes carefully crafting feedback. Quick gut reactions work better
💣 Build on sparks - When something interests you, even slightly, tell the AI to push further in that direction
🖼 Reframe it - if it feels off track, ask – is there a better question I could be asking? A better avenue to explore? The AI’s “not so good” examples can help you realize what you actually want.
Don’t settle for average when excellence is just a few iterations away. And it’s not like the AI will get annoyed with you, it will just keep going with you over and over.
Remember—be fickle and hard to please!
💡That’s your one thing for this week!
Make sure to subscribe to get the next 16 posts, every Friday morning at 5am PT/8am ET! Let’s make Friday our day to grow our AI skills experimenting on high level work challenges.
I count on word-of-mouth to grow this series, please consider sharing it with a friend.
👋Thanks,
Devin McNulty
Co-Founder Funmentum™ Labs
👀👀👀
P.S. I can get your team acting like this—book 15 minutes to pick my brain.
P.P.S. Try Funware Ideator for free, the world’s first ever automatic brainstorm, powered by AI (of course)!
🦾AI as a Creative and Strategic Partner
As we progress, each post will be linked below for reference.
Chapter 1: Mindset is the Magic
Post #4: Be Fickle and Hard to Please (this post!)