Still waiting for the AI LLM based ad autobidder so that I can just plug a machine to Google and press the "give them all my money" button.
What I see is that clients that invest in a campaign do not want to think about what an AI can produce. They don’t want to interact or brief an AI, they don’t want to do feedback rounds with an AI. They want a group of professionals that knows them to take over and do it all. If the professionals then use some AI for it, they mostly don’t care.
This is true so far for any campaign that allocates relevant funds (mid 5 figures and upwards). When it comes to the actual creation phase, right now AI is fundamentally immature and incapable of being controlled past the creation of static content.
All the motion and animation part for example is still somehow terra incognita for these tools. Take Adobe Animate, which is the go-to tool for anything 2D-animation, or Google Web Designer. Zero AI-features, simply because you can’t LLM frame by frame animations and have a result that is as precise as you need it. Or maybe you can, but for some reason these companies don’t see a business case for allocating resources to this specific development.
These tools can be great for smaller business that won’t have access to large campaigns, but as someone else mentioned, why do that when hiring a working gen-z social media native student will cost you slightly more, and possibly perform 100-times better with their native social media aesthetic?
Ps: Pomelli means door handles in Italian, and that’s… weird? Feels like a name randomly regurgitated by an LLM as well.
I know this new tool looks to be for static graphics; but I do think the same thing applies. Not using AI-generated polished graphics will become a differentiator.
If I were an EvilGoogle manager, I'd have an enshittification playbook complete with a timeline and KPIs/OKRs mapped out - and probably already linked to individual engineer's promotion/RIF futures.
They know exactly who's using this tool and which company they're using it on behalf of.
In the short term I'd have those companies webpages using Pomelli generated content to rank highly, and for advertising on those pages to show higher then usual clickthrough rates - and probably gradually downrank non-Pomelli pages on their sites. Once it becomes well known that Pomelli generated content genuinely generates more revenue that other options (even though that's only because Google have their thumb on the scale), everybody is going to jump on the gravy train, and a sub-industry of Pomelli consultancies/agencies will show up, like specialist SEO firms did way back.
Gradually that new "Pomelli Content Optimisation" will capture a significant-enough slice of the web content generation pie, and Google will start to sell them "Pro" subscriptions and features, while at the same time reducing functionality and effectiveness of the tools individuals and end-user companies have access to - driving even more revenue into the PCO industry.
Eventually, when enough companies are fundamentally reliant on external PCO vendors, Google will ramp up the pricing of their tools.
(With any luck AGI will have turned us all into paperclips before that runbook plays out.)
Guessing: because they have AI products in the pipeline that can create Youtube shorts or similar.
This aspect will be interesting to watch.
Edit: Youtube Premium should include an optional AI slop filter.
Sometimes even with a US account some things flop when you try to use them while traveling. You'd think the richass CEOs travel a lot so they would notice this problem but then you realize they never use their own products and have meat intelligence do all their shit for them anyway.
Google appears to have their AI product game together!
But to answer the question, it looks a lot like a Canva competitor.
Nano Banana alone obsoleted all of Photoshop. (And the Chinese versions of Nano Banana are even better!)
I'm most worried for my friends in creative though. I have some extremely talented friends at WPP and other agencies. Everyone is shaking in their boots.
Nobody's buying ads because of the economy, then these tools are nipping at their heels. They've already had one massive round of layoffs, and there's another one supposedly happening early next year.
Where are these millions of people going to go? These are six figure income earners.
There are five million marketing professionals in the US. If half of them lose their jobs, then what? What's lined up for them after this?
If AI fails, the economy goes boom.
If AI succeeds, the economy goes ... bigger boom?
I used to think the tools would wind up creating more work, especially in narrative creative work. Outside of A24 and indie/foreign films, Hollywood is so trite. These models drop Pixar/Disney VFX into the hands of every YouTuber - and that could be really cool when used by the right people. Like the Corridor Crew folks.
Maybe gaming and media will see a boost, but advertising and marketing folks are really going to get hit hard.
Their blog post has some detail: https://blog.google/technology/google-labs/pomelli/
Kagi said the "Key features and functionalities of Pomelli include:
Content Generation: Pomelli can generate various marketing assets
such as social posts and ad creatives by analyzing
a company's website to understand its brand identity
Brand DNA: The tool builds a "Business DNA" from a company's
website to ensure generated content is consistent with the brand's identity
Campaign Creation: It aims to generate entire on-brand marketing
campaigns with minimal user input
Editable Assets: The generated campaign assets are editable
Canva Alternative: Pomelli is positioned as a competitor
to design tools like Canva"If the vision is diluted due to lack of control afforded by AI tools, then the tools won’t be used.
Many times in Hollywood have we seen directors spend unjustifiable amounts of money in the pursuit of creative control.
Hand camera tracking a dinosaur in Jurassic Park, developing a novel diffraction algorithm for THE ABYSS, hand-drawing 3-Dimensional computer animations for 2001, creating an entire scale model practically for a single fight scene in LOTR.
AI allows you to get anything. The best movies are a direct reflection of a particular vision. AI can’t provide this and I see no way to solve it.
A natural response is - well directors already outsource some creative control to VFX artists so why not to a machine instead.
Because an artist can control everything. Even if the artist is prompting a model, at the end of the day an artist can drill right down to the tooling itself (photoshop for example) and exactly achieve the vision.
I don’t see AI achieving this granularity while maintaining its utility. It’s a sliding scale of trading utility as a time saving device for control.
If you lean too far to the control side, well you might as well fire up photoshop. If you lean too much to the utility side, you sacrifice creative control.
When looked at under this lens the utility of AI generation is actually limited as it solves a non existent problem. One can think of it as an additional piece of tooling for use only as a generational tool where there is less need for control, such as for background characters.
The team at Red Barrels, for example, train a local model on their own artwork to automatically generate variant textures for map generation. Things such as this. No need to be doom and gloom about this stuff.
You should look at ComfyUI.
Control is here, it's just not widely distributed or easy to use.
If you're patient, you can fully control the set, blocking, angles. You can position your characters, relight them, precisely control props, etc. You have unlimited control over everything. It's just a mess right now.
Oct 28, 2025
We’re introducing Pomelli, a new AI experiment from Google Labs that helps small-to-medium-sized businesses easily generate scalable, on-brand social media campaigns.
Daniel Adonai
Senior Product Manager, Google Labs
Bea Alessio
Senior Product Manager, Google DeepMind
Creating impactful, on-brand content can often require significant investment in time, budget, and design expertise. For small to medium-sized businesses (SMBs), this can be a major obstacle. That’s where Pomelli, our newest experiment from Google Labs in partnership with Google DeepMind, comes in. Pomelli is an AI marketing tool designed to help SMBs more easily generate scalable, on-brand social media campaigns to help grow their businesses.
Pomelli uses AI to understand your unique business and generate effective, tailored campaigns in just three steps.
1. Build your business DNA
Enter your website, and Pomelli will analyze it and create a "Business DNA" profile for your brand. By analyzing your website and existing images, Pomelli is able to automatically extract and understand your business’ unique brand identity. This profile includes your tone of voice, custom fonts, images and color palette.
All content Pomelli generates is grounded in this DNA, ensuring your content feels more authentic and consistent across all channels.

2. Generate tailored campaign ideas
Once your Business DNA is established, Pomelli generates tailored campaign ideas specifically for your business. This feature tackles the common pain point of coming up with fresh, strategic ideas, allowing you to quickly pick a campaign focus. If you have your own idea, you can type in a prompt to create content tailored exactly to your vision.

3. Edit and create high-quality, branded creatives
Finally, Pomelli creates a set of high-quality, on-brand marketing assets designed to help grow your brand across various channels, like your social media, your site and your ads. Browse through the generations and select the assets that best fit your campaign goals. You're in full control to make edits to the text or images right inside the tool. All assets can be downloaded and are ready to be used across your channels.

Pomelli is launching today as a public beta experiment in the United States, Canada, Australia and New Zealand in English. It’s an early experiment and it might take some time to get things right. Our goal is to build the highest quality experiments, so your feedback is appreciated. Give it a try and let us know what you think.
Related stories
.
Seems like Google will kill a whole bunch of SaaS companies with this.
Have you seen their announcements during Adobe Max? The AI features are mind blowing. Adobe is alive and well.
You have to either have some big cajones or be totally lost to think it's a good idea to create a startup that is just a simple cheap veil on someone else's extremely advanced and expensive product
What sort of market dynamics do people predict here, winner takes all? Especially when this is integrated into the platforms of distribution.
If no one uses it, that means the market has proven, no audience for this kind of product. Google loses, everyone else loses.
If everyone who wants this sort of thing uses it, that's it, Google won, everyone else loses.
The outcome to sell to investors is the least believable: people will pay for some offering when a nearly identical one is available directly from Google for free. And anyway, they have the best generative creative tech, so how could anything be better than Google's?
I wonder.
They're doing well with their existing customer base of digital creatives and related industries/professions.
Who may all be the buggy whip makers of the late 2020's.
Way too many of the people/companies who traditionally paid highly skilled and creative Photoshop users are rapidly moving away from doing that in favour of cheap GenAI slop.
I'm sure there are people in graphic design, illustration, videography, photography, UI/UX, 3D art, augmented reality, social media, creativity and design, collaboration and productivity, and education who are super excited about what Adobe is doing. I'm also sure almost all of those people are very concerned about their career choice and future (or are ignoring the reality of what's going on around them).
Sure, the top graphic designers in the world will still earn great money being highly creative for key clients. But the vast majority of people in those fields are not the top in their field, and the vast majority of clients those people invoice are going to consider cheap AI slop "good enough" for their businesses and use cases.
I have a 30+ year career in web related roles, working more or less closely with graphic designers, artists, illustrators, photographers, and other website development related professions. All of the ones I've remained friends with over that time are either deeply concerned about their career future, or have already jumped ship and become nurses, carpenters, teachers, caregivers, and even priests and drug dealers...
It doesn't look like they developed any models. The 3d relighting and 3d manipulation are all 3rd party models given a UI.
As a simple example, assume a specific LLM-based tool (like Google's own, or someone else's) happens to generate a social media mascot for you that looks a lot like the modern rendition of Mickey Mouse.
Let's see how long that creation flies as public domain because it came out of an AI (that almost certainly consumed a giant amount of content produced by Disney as part of its training).
If you want a specific tool, here is Elsa with a cigarette generated using Midjourney and more: https://journeyaiart.com/tag/Elsa .
Aside from the last one, that kinda sounds like a win for society.
See also [1] mentioned in the framework linked by sibling comment, AI copyright is essentially a logical extension of this.
[1] https://en.wikipedia.org/wiki/Monkey_selfie_copyright_disput...
But what happens if they MIX some of their own code with AI-generated code, is that combination then their copyright? With such combined output it would be very difficult to determine which part was created by human, which by AI, and which by AI but slightly modified by human.
In the domain of graphics the AI could put in some markers which tells the graphic is AI-generated, but with code that is probabaly not possible, code is code and can always be edited by humans.
A separate question is that if I use Claude to generate some code but then stamp the output with my copyright notice, am I doing something illegal?
The short version is the copyright office says it is possible works by creative human authors using AI tools are partially copyrightable in many cases.