Dorian Dargan and the Rise of Photorealistic AI Shopping

Online fashion has always asked shoppers to imagine – but imagination is precisely what causes doubt, disappointment, and returns. Dorian Dargan, co-founder of Doji, is replacing guesswork with photorealistic simulation by using diffusion models to let people see exactly how clothes will look on their own bodies. This article explores how his work is turning AI into a visual decision engine – and why it could redefine the future of digital commerce.

Table of Contents

Key Takeaways

AI is shifting e-commerce from browsing to simulation. Doji turns shopping into a visual, personalized experience where customers see outcomes instead of guessing.
Diffusion models enable photorealism, not just approximation. The technology allows clothing, bodies, lighting, and materials to be rendered in a way that feels photographic.
Reducing returns is only the first-order benefit. The deeper impact is increased customer confidence and better decision-making.
Personal avatars are becoming a new interface layer. Doji points toward a future where your digital self becomes a persistent way to interact with products.
The real innovation is decision transformation. By turning AI into a mirror instead of a recommender, Doji changes how people choose.

Dorian Dargan: The Future of Online Shopping Isn’t Browsing – It’s Trying On

For two decades, e-commerce has asked consumers to make a leap of imagination: to believe a flat image on a screen can represent how something will look on their body, in their life, and in motion. That leap has been expensive. Fashion e-commerce still suffers from some of the highest return rates in retail, driven largely by fit, sizing, and expectation mismatches.

Dorian Dargan, co-founder of Doji, is building a different future: one where online shopping is no longer an act of guesswork, but a visual, personalized, and photorealistic experience. By using diffusion models to generate high-fidelity avatars and virtual try-ons, Doji is turning AI into a mirror – one that could fundamentally change how people buy clothes online.

View this post on Instagram

A post shared by Dorian Dargan (@doriandargan)

Problem: The Structural Problem with Digital Fashion

The core limitation of online fashion has never been logistics. It has been representation.

Clothing is experiential. Fit, drape, proportion, posture, and movement matter as much as color or price. Yet the digital fashion industry has been built on static product photos, standardized models, and size charts that assume bodies behave like geometry.

The result is a system optimized for browsing, not confidence.

Return rates in online fashion routinely exceed 25–40%, creating massive costs in logistics, sustainability, and customer frustration. Brands spend heavily on performance marketing only to lose margin in reverse logistics. Consumers learn to “order three sizes and return two,” turning inefficiency into habit.

For years, the industry tried incremental fixes: better size charts, more angles, video, influencer photos. But these still rely on abstraction. None answer the real question a shopper is asking:

“What will this look like on me?”

Dargan’s insight is that this is not a UX problem. It’s a representation problem.

Innovation: Turning Diffusion Models into Personal Fitting Rooms

Doji’s core technical leap is using diffusion models – the same class of models behind modern image generation – to create photorealistic, body-faithful avatars and realistic virtual try-ons.

Instead of placing clothes on generic models or warping garments onto static photos, Doji builds a personal visual simulation of the user. The system learns how fabrics behave, how garments sit on different body types, and how lighting, posture, and perspective affect perception.

The result is not a cartoonish approximation. It’s a high-fidelity visual prediction.

What makes this approach different from earlier “virtual fitting room” attempts is realism and generality:

The avatar is personalized, not templated
The clothing is rendered, not pasted
The output looks photographic, not illustrative

Under the hood, diffusion models allow the system to reason about texture, shadow, shape, and material behavior – producing images that feel more like photography than simulation.

In effect, Dargan is not building a sizing tool. He is building a visual decision engine.

Why This Is a Real Platform Shift – Not a Feature

Most e-commerce tools try to optimize conversion. Doji is changing how decisions are made.

Instead of:

“Do I like this product photo?”

The question becomes:

“Do I like how this looks on me?”

This is a subtle but profound shift. It moves shopping from catalog browsing to personal visualization.

Just as Google Maps changed navigation from instructions to simulation, and Airbnb changed travel from hotels to lived-in spaces, Doji is changing fashion e-commerce from representation to embodiment.

And because the system is generative, not manual, it scales across different bodies, garments, brands, and aesthetics.

This is what makes it infrastructure, not a gimmick.

Impact: Fewer Returns, Higher Confidence, Better Data

The immediate business impact is obvious: return reduction.

When customers can see realistically how something fits and looks, they buy with more confidence and return less. That alone has massive implications for logistics costs, sustainability, inventory planning, and brand margins.

But the deeper impact is behavioral.

Doji changes how people explore style. Customers can try bolder choices with less risk. Brands can present more diverse bodies without reshoots. Personalization becomes visual, not just algorithmic.

For brands, this also creates a new layer of data: not just what customers click or buy, but what they try on, compare, and visualize.

That turns fashion from a guess-based funnel into a simulation-driven decision system.

Why This Matters Beyond Fashion

Dargan’s work sits inside a much larger shift: AI as interface, not just engine.

For years, AI improved search ranking, recommendations, pricing, and ads – but the user experience stayed the same.

Doji represents a different direction: AI that changes what the user sees and how they decide.

This pattern will not stay in fashion. The same approach will apply to furniture and home design, eyewear and cosmetics, fitness and body modeling, and even medical visualization.

The deeper idea is this: generative models are becoming tools for personal simulation. And simulation is the highest form of decision support.

The Founder’s Role: Bridging Research and Reality

Dorian Dargan’s role is not just technical. It is translational.

Many diffusion breakthroughs remain stuck in labs or novelty apps. The hard work is turning them into reliable systems, fast pipelines, scalable infrastructure, and brand-safe outputs.

Doji’s innovation is not just that it can generate images – it’s that it can do so consistently, predictably, and usefully inside a commercial workflow.

That is the difference between a demo and a platform.

Future Outlook: From Shopping Tool to Visual Identity Layer

In the long run, Doji’s technology points to something bigger than shopping.

As digital identity becomes more visual across social, work, gaming, and commerce, personal avatars will become persistent representations, not one-off tools.

Your “Doji self” could try clothes, appear in brand experiences, show up in virtual environments, and become your visual interface to digital commerce.

In that future, shopping is no longer about products. It’s about you in different worlds.

Dargan is not just fixing e-commerce conversion. He is helping build the visual layer of personal AI.

When AI Stops Recommending and Starts Showing

Most AI systems tell us what to consider. Doji shows us what’s possible.

That difference – between suggestion and simulation – marks a real shift in how humans will interact with machines in commerce. By turning diffusion models into mirrors instead of toys, Dorian Dargan is helping move retail from abstraction to embodiment.

In the next era of digital shopping, the most important interface won’t be a feed.

It will be you.

FAQs

Who is Dorian Dargan?

Dorian Dargan is the co-founder of Doji, a company building AI-powered photorealistic avatars and virtual try-on technology for fashion e-commerce.

What is Doji?

Doji is a retail technology company that uses generative AI to let users try on clothes virtually using personalized, photorealistic avatars.

How is Doji different from older virtual fitting room tools?

Earlier tools relied on static templates or simple overlays, while Doji uses diffusion models to simulate realistic bodies, fabrics, lighting, and fit.

Why does this matter for e-commerce businesses?

Because fashion has extremely high return rates due to fit uncertainty, and Doji helps customers buy with confidence.

What is the long-term vision?

Personal avatars could become a persistent visual identity layer for commerce, social platforms, and digital environments.

Sources:

Photo credit: Dorian Dargan / Instagram