Table of Contents
Hey, what’s up, hello.
Coming to you on Monday, as we were off on Friday for the long weekend!
Just as a quick aside, if you are feeling overwhelmed with all these damn tools out there. YOU ARE NOT ALONE!
Every week, it feels like there’s a new AI tool that’s supposed to change the game.
You may have learned about Flux before Nano Banana appeared.
You were still making peace with Veo 3 before Veo 3 dropped with frames.
And then Meta and Midjourney decided to start holding hands (cuties… well, sorta).
If you’ve ever felt like your “workflow” has the shelf life of an avocado (low key don’t hate brown avocados), you’re not alone.
Or you have no idea what I am talking about, and that’s why you are here!
Welcome!!
Lesson: The Stack Beats the Silver Bullet
Here’s the real play: there won’t be a single winner (at least for a while)… so keep the below in mind:
Meta + Midjourney aren’t explicitly coming for ads… yet. So use Midjourney to animate all of your statics in the meantime!
Google will own the net new static scene given its font and product fidelity with Nano Banana
Veo 3 is where you lock in motion + product fidelity
Combining Nano + Veo 3 (workflow coming soon) will be the move
Meta x Midjourney
Meta isn’t dipping a toe; they’re Ron Burgundy cannonballing into the AI creative pool. Llama might suck, but with Midjourney plugged in, the creative firehose is coming.
Zuck makes a splash!
Here’s what to know:
Meta isn’t partnering with Midjourney to build an ad-specific framework or a Vogue-level creative studio, at least not yet.
It’s intended to upgrade Meta’s multimedia AI, encompassing everything from feeds in the Meta AI app to image-generation features in Facebook, WhatsApp, and Instagram.
Translation: your average user will start pumping out Midjourney-quality content natively inside Meta’s ecosystem.
TL;DR: It is more user-focused than advertiser-focused… for now.
Let’s talk b-a-n-a-n-a-s, 🍌 Gwen Stefani style: (Google’s newest model)
(Okay, real talk, without that song, would any millennial be able to spell banana? Yes? Okay, cool, so it’s just me. Word.)
Banana makes other image models look like dial-up internet*: faster, cheaper, better humans. And it gets fonts right. (Do you know how many hours of my life I’ve lost to AI spitting out Comic Sans knockoffs? Too many.)
*Side note: Did you know AOL officially shut down its dial-up service last week? Tragic.
Being able to upload multiple images is really the powerhouse of the model.
And despite Twitter telling you otherwise, it does NOT have a video component. It is strictly an image model.
The tweet from Teejay below combines Nano and Veo 3.
USE CASE 1: Cloning competitor’s ads using your own fonts (ooooh, ahhhhhh)
I cloned a Manscaped ad, swapped in a Go Brewing can, and Banana matched the font hierarchy perfectly. It felt… illegal. (All of my info below so you can judge me.)
It did take a few attempts, but I am happy with the end result (perhaps minus “Moisturizer” lol).
Outcome:
Reference
Output
Images uploaded:
Prompt:
Make a 1x1 image
Swap out the Manscaped product for the attached Disarm Hazy IPA can.
Replace the "Manscaped" headline with "Disarm".
Replace "Crop Preserver" copy with "Flagship Hazy".
Replace the USPs exactly as follows, keeping their same placement and font size from the reference:
- Replace "Deodorant & Moisturizer" with "Juicy, full-bodied"
- Replace "Long-Lasting" with "Notes of pineapple and mango"
- Replace "Advanced, Quick-Absorbing, Clear Drying Lotion" with "Non-alcoholic"
Follow the font hierarchy and layout style of the attached INTRODUCING DISARM example:
- Large headline at top (“Disarm”)
- Medium subheadline under product name (“Flagship Hazy”)
- Supporting USPs with icons arranged around the can in the same positions as the Manscaped referenceUSE CASE 2: Creating net new content
Nano Banana is the first model that excels in product fidelity with people.
For this one, I also had Veo 3 animate the Banana image with an extremely simple prompt to showcase the power of the two models together.
Outcome:
Nano Banana
Animated with Veo 3
Images Uploaded:
Prompt:
Generate a photorealistic scene of a fisherman sitting on a small boat, casually casting a fishing line into the water. Place the branded cooler on the boat, open and filled with ice and the same branded beer cans. Ensure the cans appear high fidelity, upright, and not warped. Add the craft beer award badge above the scene without blocking any elements. Finally, include one oversized version of a beer can, positioned naturally in the image (like a product spotlight) to emphasize label clarity and fidelity. Lighting should match a sunny day on the waterVeo 3 Product Fidelity
Now, full disclosure: this workflow is basically the same one I dropped on LinkedIn on Wednesday.
So if you already read that and thought “wow, this guy really cracked the code,” congrats, you’ve seen behind the curtain and realized the Wizard was just a goofball the whole time.
However, since not everyone wants to doomscroll on LinkedIn (a healthy choice, by the way), I’ll repeat it here.
Yes, it might get stale as Nano Banana evolves, but right now this is still the cleanest way I’ve found to keep product labels sharp in Veo 3 without them turning into melted wax:
👉 Notion Doc Here 👈
That’s all for this week!
I imagine that this week, I will be spending a lot more time on Nano Banana and combining it with Veo 3 for workflows.
I hope to build an n8n flow I can share as well.
Additionally, I’ve had some questions about consulting to teach in-house teams how to navigate all of these workflows.
Let me know if folks might be interested in something like that, and I’d be happy to chat.
Have a great week!
Will