AI agents for fashion are the next step after the image-generation tools we have all got used to. In September 2026, the US company Raspberry AI expanded its platform into a single “agentic” system that follows a garment from the first moodboard all the way to the product page of an online shop. Among its clients it names SKIMS, ALO, Savage X Fenty, Beyond Yoga, Oscar de la Renta and Walmart.
In this article we explain, simply: what is an AI agent? What does this platform actually do? What does the much-quoted 75% figure really mean? And most importantly: what does it mean for you as a designer?
What are AI agents for fashion?
An ordinary AI tool gives you one result when you ask: you type a description and get a picture of a dress. An agent is software that carries out a chain of tasks on your behalf, passing what it learned at each step on to the next. Picture a studio assistant who doesn’t just sketch for you, but gathers trend images, builds the moodboard, suggests fabrics, writes the technical specification, and prepares the catalogue photos — while you review and approve at every stage.
That is what AI agents for fashion offer, and what Raspberry AI’s CEO, Cheryl Liu, meant when she told FashionUnited the shift is from “isolated outputs to connected work”.
What did Raspberry AI announce?
Raspberry AI is a New York start-up founded in 2022. In January 2025 it raised $24 million in a round led by Andreessen Horowitz. On 16 September 2026 it announced that it had brought AI agents for fashion together in one workflow covering five departments inside a brand:
- Design: trend research, moodboards, fabric exploration from the brand’s own library, and design development.
- Merchandising: building the range and showing it on models before a single piece is made.
- Wholesale: buyer presentations and lookbooks.
- Marketing: campaign imagery.
- E-commerce: product images and product pages.
Liu also mentioned the tech pack — the file a designer sends to the factory with measurements, materials and construction details — which can now be generated in the platform and sent straight to manufacturers. A partnership with Pantone keeps the official colour code attached to a design from first idea to supplier, which reduces colour mismatches between sample and production.
The 75% figure: what does it really mean?
The company says brands using its platform see 2–5 times faster speed to market, 60% lower sample costs, 80% lower photoshoot costs and 75% lower production costs.
But there is an important point to keep in mind: these are figures reported by the company itself about its clients in general. They are not published results from SKIMS or Beyond Yoga specifically, and the company has not explained what baseline it measured against. Analysts have also pointed out that software cannot lower the price of fabric at the mill or the cost of sewing on the factory floor. So the savings most likely sit in the development stage: fewer physical samples, fewer photoshoots and less time lost on revisions and back-and-forth.
What this means for you: whenever you read a shiny number in tech news, ask who said it and what it was measured against. That habit protects you from buying or investing decisions built on marketing.
Why are big brands paying for AI agents for fashion?
The business motive is clear. In fashion, range decisions are made many months before a garment reaches the customer, and every piece goes through dozens of steps: research, sketch, material choice, technical design, sample, fitting, revision, approval, production and photography. Every physical sample costs fabric, shipping and time.
The bigger problem is stock that never sells: Raspberry AI estimates that around 40% of apparel produced each year goes unsold or is heavily discounted (the company’s own estimate). We looked at the scale of this problem in Excess Inventory in Fashion.
Liu’s argument is a smart one: the goal is not more collections, but a fashion calendar that is “more responsive, but not necessarily more frantic”. A brand can delay its final decision and make it with fresher information about what customers want — and so produce what will actually sell.
Where does the designer fit in alongside AI agents for fashion?
This is the most important part of the story. Liu herself says people set the brief, define the customer and choose which trends suit the brand. Designers stay in charge of silhouette, proportion, fabric, washes and detailing. She admits AI doesn’t know on its own which idea a brand should back, and cannot judge fit from a screen — so physical samples and fittings remain essential.
The company even hired in-house designers and merchandisers to teach the software the language of the trade: “fuzzy sweater”, for example, refers to a specific yarn texture, not just a look. In other words, professional expertise is what makes these tools work in the first place.
What does this mean for you as a designer?
- Learn to write a good brief: AI agents for fashion are only as good as your instructions — who is the customer, what is the occasion, which fabric, what budget?
- Build your own library: big platforms work from a brand’s own fabric and colour library. Start photographing and organising your materials now; they are the memory of your designs.
- Keep your pattern and fitting skills: as long as AI can’t judge fit on screen, pattern cutting and the toile are what set you apart.
- Understand the tech pack: if tech packs are generated automatically, the designer who can check them and spot the errors is the one with the advantage.
- Start with digital sampling: the way into this world begins with tools like CLO 3D, as we explained in Digital Sampling in Fashion.
The takeaway
AI agents for fashion are not replacement designers; they are assistants that take over repetitive admin so designers have more time for creativity and decisions. Company-reported numbers always deserve a critical read, but the direction is clear: big brands want to design faster, produce less and sell more. At Anna Stella Academy we cover these changes in our AI in Fashion unit, so you are ready for them with real skills, not just headlines.
Read also
- Agentic Commerce with AI: How Did Zara and H&M Capture ChatGPT’s Visitors?
- AI Product Photography
- Excess Inventory in Fashion
Sources
