“Claude Beige” Is the New Stock Photo
- Amy Winner

- 4 days ago
- 4 min read
You know stock photos? The ones with the suspiciously happy team around a conference table, or the woman laughing alone with her salad. They’ve been a staple shortcut for marketers, but also made everything look like everything else. We trained ourselves to say they were a realistic compromise to match the available budget and timeframe. Then came tools like Canva and Nano Banana, and suddenly you could easily generate unique images that matched your brand guidelines. Slowly we’re seeing the (welcome) extinction of the laughing salad woman.

While stock images are slowly becoming extinct, we have a new replacement: “Claude Beige.”
Over the last six months, Claude Beige has crept into assets that clients send us as we’re working together. From decks to homepage drafts to new page wireframes, we’re fielding mockups built in ChatGPT, Claude, and Gemini all the time. Sometimes, we get all those things, plus a full campaign brief, delivered on a Monday morning with some "took me about 20 minutes to pull this together so we must be able to launch by Thursday!" energy. I say this jokingly, but also not really.
Here's the thing: We can all spot the AI-generated stuff immediately. Not because the product that AI puts out is bad, but because it's safe, boring, and consistent. The wireframes all use the same three-column hero layout, the copy is full of em dashes (RIP em dash, you were my favorite punctuation that, as a result, I can never use again), sentence fragments, and groups of three. The messaging sounds like it was written for every company in that business category at once, because in a way, it was. It is the marketing equivalent of a stock photo: technically functional, instantly forgettable.
Coincidentally, my co-founder Elise Oras just (separately) wrote a blog post about “the beigeification of the internet” and sent me the link to read along with an “LOL,” when she was reviewing this newsletter draft. She digs into the backstory of the AI training data origins that led to this (very documentable) phenomenon. Give it a quick read if you want to geek out with us.
We love AI at Wheels Up. We use it every day. We recommend it to clients. We've built parts of our own workflow around it. It is a tremendous time saver and helps us deliver more (and better) marketing for our clients at the same price point. The question isn't whether to use AI, it’s whether you use it as a starting point or as a finished product. Right now, we are seeing so much (published!) marketing that’s just AI slop. So, it’s worth thinking through the right application of these new tools.
Where AI earns its keep
AI is genuinely very good at high-volume, short-shelf-life work: SEO pages, brainstorming sessions, first-pass wireframes, reorganizing a pile of information into a usable structure. These are tasks where the goal is coverage and volume. The output doesn't need to sound like you, it just needs to exist. For that work, AI is fast and good. It’s also a great generator of the outline or rough draft. Especially when you’re having the hardest time just getting started, it can be a great tool to give you something to sink your teeth into.
The problem starts when the work you’re doing actually needs to sound like you: website homepage, investor deck, brand messaging, a critical email that’s going to your 40 most important customers. That work requires a voice, a point of view, and an understanding of your specific market that no model (even a custom GPT or Claude Project) can have yet.
What you lose in the copy/paste
There are two big risks in going straight from Chat to publish. First is errors. Not hallucinations, but actual nuanced inaccuracies about the product or mismatched market understanding and vernacular. Your team knows the nuanced stuff. AI does not.
Number 2 is connection. People are especially hungry for authenticity right now. The brands that are winning are the ones that sound like humans made them, because humans did. Every piece of content you put into the world is a chance to say something specific, or to make your audience feel like you actually understand their problem. AI will give you a sentence. You give it meaning. As attention spans get (even) shorter, AI gets (even) better, and budgets get (even) tighter, I don’t know how long this opportunity to be truly authentic in your marketing will last.
It keeps me up at night
I own a marketing agency. I think a lot about where this is all going. Will agencies exist in 5 years? In 3 years? In 2027?? Lucky for us, I don’t think we’re at risk of imminent extinction. I believe there’s still a good chunk of human-generated marketing work that survives because it does require sound judgment.
But I do think we are experiencing a lot of companies and marketers whole-heartedly farming out this strategic work to AI. I think sometimes AI is used to fake expertise that’s lacking, or to shortcut the hard work of really getting to know the market. And those are two areas that can cause catastrophic ripples down the line.
I get it. Marketing is hard. Growing a business is hard. But AI doesn't change that. It’s still going to be hard. What AI can change is how you tackle the hard. Offload the AI-able tasks and spend the time on the work that only you can do—there's still a lot of it. (One last em dash for old time’s sake…)
How we’re thinking about it at Wheels Up
At Wheels Up, we're not standing here watching AI happen. We’re embracing it and (hopefully) making smart choices about it. We're AI practitioners, doing the in-the-trenches evaluation work so our clients don't have to. We are deciding very intentionally which tools to use, for what tasks, and where in the process. And we always hand it back to a human for (at the very least) a final QA to protect the quality of the work and the integrity of the brand it represents.
If you want a second set of eyes on how AI fits into your current marketing workflow, or you want to walk through your content process and figure out where you're leaving human value on the table, let's talk. Grab time on my calendar. This is exactly the kind of conversation we're having with clients right now.




