← Back to blog
|

How I Find AI Art to Sell on Etsy Before It’s Saturated

My research process for AI wall art: spot a trend upstream on Pinterest, validate it with Etsy data, build a mood board, then prompt. Real example.

If you’re making AI art for Etsy, here’s the one thing I’d stop doing immediately: opening Midjourney or ChatGPT and making artwork just because you like it. That’s one of the easiest ways to end up with hundreds of listings and very few sales.

 

 

 

Instead, I’ll show you the research process I use when a new trend signal appears, with a real example taken all the way from trend report to finished artwork. For context, I’ve sold over 5,500 pieces of art on Etsy in less than a year, so this is what I do, not theory. Some examples here sit outside what I’d sell in my own shops on purpose: I want you to see how the thinking transfers.

The mistake: waiting until a trend shows up on Etsy

By the time something is obviously trending on Etsy, hundreds of other sellers are looking at the same numbers you are. If your plan is “vintage landscapes are selling, I’ll make vintage landscapes”, thousands of sellers had the same idea, and the market’s flooded before you’ve listed one piece.

So I don’t start on Etsy. I start further upstream. And once I have a direction, I want to answer three questions:

  • Who is buying this?
  • What are they actually buying?
  • Why would they choose my version over someone else’s?

Look at hobby reports, not home decor reports

This took me a while to figure out. The best place to spot a wall art trend isn’t a home decor report. It’s a hobbies report.

Think about how people decorate. They buy art that reflects who they are, what they’re into and what they spend their weekends doing. When a hobby spikes, the decor that reflects it usually follows, often before it shows up in a wall art or home decor report. By the time it hits a decor report, it’s already mainstream.

A real example: Pinterest’s 2026 hobbies report

The week I filmed this, Pinterest released its 2026 hobbies trend report. It covers nine areas, including Kitchen Witch, Cosmic Reset, Backyard Botanist, Weekend Equestrian, Curated Clutter and Casual Crafter.

Under Casual Crafter, in a cluster called Stitched and Sculpted, searches for hand embroidery were up 1,351%, alongside clay bag charms and beaded bag charms. (The report is only available in some regions; select English (US) in the Pinterest Newsroom to view it.)

 

Why focus on embroidery? Not because of the big number. Because people are already spending real money on it.

Step 1: Validate it with Etsy’s own search data

You could use Google Trends or EverBee for this, but here I used Etsy’s own Marketplace Insights tool, because it shows what buyers actually type before they look at listings.

“Embroidery” showed over 74,000 searches. It was flagged as low conversion, but the graph showed a real seasonal climb to about 15,000 monthly searches by July. Etsy’s data was backing up what Pinterest was seeing.

Step 2: Check that people are actually spending

A spike doesn’t pay the bills, so next I looked for real sales. One embroidered baby sign had sold an estimated 1,850+ times (EverBee’s estimate), with around a quarter of a million views and nearly 16,000 favourites, at $52. It’s physical, handmade and personalised. Another listing carried Etsy’s bestseller badge at over $31. Two price points, same buyer, real money.

Step 3: Work out what buyers are really responding to

Nobody searching “embroidery” necessarily wants an embroidery hoop on their wall. They’re responding to a look: stitched, tactile, handmade, a little imperfect, nostalgic.

That look can sit on completely different subjects, which means it can reach completely different buyers. And there’s a buyer who wants that feeling at a more accessible price point. That’s where digital wall art comes in.

So the real question isn’t “what’s my niche?” It’s “how do I translate this look for a more price-sensitive customer across different subjects?”

Step 4: Find the look in other formats

Next I searched for the aesthetic in digital form. There are printable downloads literally titled “embroidery style”. None are physically stitched, and they sell across a wider price range than the handmade originals. Buyers are already accepting a printed, illustrated version of the look. It doesn’t need to be embroidered. It can be made with AI.

Step 5: Build a mood board

I pulled the best examples onto a mood board to see what repeats. Here it was muted vintage palettes (navy, cream, rust, olive), flat stitch textures and small folk motifs. That’s the visual language I’m aiming for.

Step 6: Turn the mood board into prompts

I don’t sit down and write prompts from scratch. I drop a screenshot of the mood board into a custom GPT called Promptsmith Studio and ask it to write prompts for bestsellers based on that inspiration. You could use regular ChatGPT or Gemini too.

I didn’t build Promptsmith Studio. It comes free with a community I’m in, the AI Art Sellers Collective (free group).

It came back with 10 prompts. Here’s one, so you can see the structure:

A whimsical western folk embroidery sampler. Cowboy riding a horse through desert cacti. Terracotta, mustard, sage and dusty blue thread palette. Primitive stitched shapes. Rustic framed textile art. Nostalgic Americana composition. Small decorative stars and geometric motifs. Warm aged linen background. Vintage folk embroidery artwork.

The result: three pieces in 15 minutes

From building the mood board to having three finished pieces in front of me took about 15 minutes. That still gets me every time.

  • The dark-humour piece. My first bet. It plays in almost the same space as a bestselling listing I’d already found, with the same humour, execution style and price potential.
  • “Our Happy Home”. Not tied to one joke or one holiday. A warm, sentimental piece that can sell steadily for years, not just one good month.
  • The cowboy. A strong visual story, nostalgic and on trend with the western pieces coming through on Etsy right now.

That’s not luck. It’s what happens when research comes first.

The full framework

  1. Go upstream of Etsy before a trend is obvious.
  2. Spot a real signal, and know why you’re picking it.
  3. Test it against Etsy’s own search data.
  4. Check that real listings prove there’s spend.
  5. Analyse the style and subjects buyers respond to.
  6. Scan other formats for the same look.
  7. Capture it on a mood board.
  8. Generate prompts, and cut the ones that just repeat what’s already selling.
  9. Test.

Until you’ve done all that, it’s a hypothesis. The goal isn’t to predict what’s guaranteed to sell. It’s to stack enough evidence behind an idea that it’s worth testing.

Once you’ve got artwork, the next step is getting it found. My process for that is in how I write Etsy titles and tags.

Where I’d look next in this report

If I were going deeper into the Pinterest report, these are the five I’d explore:

  • Backyard Botanist: pressed flowers and botanical illustration
  • Cosmic Reset: astrology and tarot
  • Curated Clutter: I think there’s something in recipe cards
  • Kitchen Witch: smaller niches sometimes leave gaps for new sellers
  • Weekend Equestrian

If you’re new to AI art and want the setup side, my AI Art Starter Guide covers tools and costs, and there’s more on my free resources page.

FAQs

How do I find trending AI art to sell on Etsy?

Start upstream with hobby and lifestyle trend reports like Pinterest’s, then validate the idea with Etsy’s search data and real bestselling listings before you make anything.

Is Etsy Marketplace Insights free?

It’s Etsy’s own tool for seeing buyer search data, and it’s what I used in this example. [CHECK CURRENT ACCESS BEFORE PUBLISHING] EverBee and Google Trends are alternatives.

How do I write prompts for AI wall art?

I build a mood board from real bestsellers first, then ask a prompting GPT (I use Promptsmith Studio) to write prompts from a screenshot of the board. The prompt should cover subject, palette, texture, composition and style.

How do I avoid making AI art nobody buys?

Don’t create from personal taste alone. Check who’s buying, what they’re buying and why they’d pick your version. If you can’t answer those, keep researching.

How long does trend research take?

The research is the slow part. Once I had the mood board, getting to three finished pieces took about 15 minutes.

Similar Posts

3 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *