Google launches AI photoshoots.

Everyone cheers.

“Finally. Infinite content. No budget. No shoots. No stress.”

You think this solves your content problem.

It doesn’t.

It makes it worse.

We ran an experiment against the robots. Let me show you.

This edition of supergoods is proudly brought to you by Stickybeak 

Consumer feedback for FMCG brands.
Stop guessing. Start testing.

Consumer goods move fast (it’s in the name).
Technology moves faster.

Research used to take weeks.
Now it takes hours.

Stickybeak gives you the confidence to decide, now.

The answer to all your problems.

Last week, Google Labs went viral on Twitter with endless AI bro's declaring this is the death of agencies.
Pomelli is Google’s AI tool designed for businesses to “quickly create on-brand marketing campaigns and content”. 
It’s what we all want right - fast and free.

But I don’t think it’s going to solve any of your problems.

So I’ve put it to the test. We've recently completed a full-scale production photoshoot for one of our clients. Everything from conceptual ideation, pre-production planning, multiple locations and multiple models through to days of post-production editing.

It was by no means free, but it probably cost less than you think.

I'm going to use the same brief, the same art direction, the same intent to recreate this photoshoot using Google Pomelli. Hell, I’m even going to give it our photos as a reference to see if it can simply iterate on what we’ve already done.

The end result - your ‘human' reference point

OK before we get into the AI stuff, here’s a snapshot of what we actually created - the end result was something like a few hundred photos but it's easily a full summer season worth of content for our wonderful client, Bar None.

Yes, we hired a Mustang - it only cost $265 for the day. Somehow we managed to find one in the brand’s hero colour.

The concept for the shoot was around the idea of life's daily moments - insignificant glimpses into a normal summer day where the product fits in seamlessly.

We were going for a 90s retro kinda-vibe.

Beach, pool, laundromat. You get the idea.

So how does Pomelli compete?

First up, it did an exceptional job at reading the brand DNA from the website. All you have to do is drop in the link and voila, it pulls apart your brand book in an instance.

Next, you upload a photo of the product you want to run a "Photoshoot” for and the magic begins.

This was the result.

I’m not just underwhelmed, but I’m also annoyed that it flipped the product around the wrong way so a quarter of the output is useless.

I felt like the woman looked too much like a character from Pluribus (read: mindless zombie) so I asked for some adjustments.

I guess she has all her fingers but the product is way out of proportion.

At this point, I felt like Pomelli wasn’t going to give me the vibe I wanted to so I'm gonna switch gears to ChatGPT.

How about Sora instead?

Frustrated with Pomelli, I’m heading to Open AI instead.

I’m giving it a decent shot by starting with a request for a prompt.

It came back with a thousand characters of word salad that I didn’t read but I fed it back into the thing to do its job. It’s taking a seriously long time to render the visual so hopefully that means top tier effort.

Here’s what came back.

Visually it nailed the vibe. It’s even got the car in the BG, nice touch.

But it has a distinct feeling of “wtf, I can’t use that”. Maybe it’s the odd styling, the giant product or the strange composition. I really wanted talent in the visual too, so I’m gonna go again.

Who wore it better?

You can’t fake effort

This isn’t some boomer take about “hard work” but the reality is that effort compounds.

When you shoot for real with real people, ideas swirl around your brain and things come together in surprising and interesting ways. You develop the style as you go, building taste and solving problems.

That compounding process is what gives content depth.

If you use the same AI tool as everyone else…
You are training on the same references as everyone else…
You prompt in the same way as everyone else…

You will converge. And convergence is death.

Content is not a game of “good enough”. There is almost zero marginal benefit in flooding feeds with average. But there is enormous upside in being unmistakable.

When production is free, attention becomes expensive

Our feeds are being flooded with this content. Right now, it’s mostly eCom kids trying to build a business on a buck, but as real brands begin to integrate this stuff into their workflows, consumers will become harder and harder to reach.

AI hasn't raised the bar and made content more accessible, it's simply lowered the cost of entry and lifted the standard for what’s worth our attention.

I’m not anti-AI but there are a load of downsides and that’s not to mention the ethical and environmental concerns.

I think we’re going to see more brands invest in craft and people. I have no doubt plenty of agencies are going to be displaced because of AI. But I don’t think it’s the death of agencies. I think it's the birth of something new. Lean, niche-specialised and fiercely independent. That’s what we’re building at Mind Control anyway.

If you wanna know what a shoot like this Bar None production cost, reply to this email.