A measured conversion lift moved this from pilot to platform
LPP, the European fashion group behind Reserved, Cropp, House, Mohito, and Sinsay, has entered a strategic partnership with Krakow-based startup Pirxe to scale spatial-AI product content across its marketing. The decision followed a pilot under the Sinsay brand that delivered 21% higher conversions and a 21.6% increase in click-through rate. Those are the numbers that turned an experiment into a rollout, and they are what a marketing-technology leader should anchor on when a vendor pitches generative content as a production replacement.
The framing from LPP is explicitly operational rather than experimental. Marcin Piechocki, LPP Management Board Member and Managing Director of Sinsay, said the group "invest in technologies enabling faster, more effective operation on larger scale." That is the language of a company treating AI content as infrastructure, not a campaign gimmick. When a retailer moves from a single-brand pilot to a group-wide partnership on the strength of measured lift, the interesting question shifts from whether the technology works to how it changes the cost structure of creative production.
Digital twins turn one asset into an unlimited variant library
The mechanic behind the results is the digital twin. Pirxe creates photorealistic 3D representations of products, which lets creative teams generate and test many advertising variants without commissioning a new photo or video shoot for each one. In fashion, where assortments turn over quickly and every product needs imagery across multiple channels and formats, the traditional model of shoot-then-edit is a persistent bottleneck on both speed and cost. A reusable 3D twin collapses that bottleneck by decoupling variant creation from physical production.
For a technology leader, the strategic value is the shift from a fixed asset to a generative one. A photograph captures a product once in one context, while a digital twin can be re-lit, re-staged, and recomposed for any channel or audience on demand. That capability is what makes rapid variant testing feasible, and rapid testing is what produced the measured lift. The twin is not just a cheaper image; it is the substrate that makes continuous creative optimization operationally possible at the scale a large fashion group requires.
Message alignment is the mechanism behind the numbers
The reason variant generation translates into conversion lift is alignment. When creative teams can cheaply produce and test many versions of an ad, they can match the message far more precisely to each audience, channel, and moment rather than settling for a single asset stretched across everything. Michal Niedziolka, Omnichannel Marketing Director at Sinsay, tied the result directly to that mechanism: "Better message alignment directly translates into higher customer engagement."
That causal chain is what a marketing-tech leader should test against their own funnel. The 21% conversion and 21.6% CTR gains are Sinsay's numbers under Sinsay's conditions, and they are not a guarantee for a different brand, catalog, or media mix. The transferable insight is the mechanism: cheaper variant production enables tighter targeting, and tighter targeting drives engagement. A leader evaluating this should ask whether their bottleneck is genuinely creative production cost, because if it is, the same lever plausibly applies, and if it is not, the twin technology will not fix a different problem.
Scale is exactly why the economics change
The context that makes this partnership serious is LPP's footprint: 47 markets, more than 3,800 stores, and roughly 63,000 employees. At that scale, the number of product images required across brands, languages, channels, and formats is enormous, and the marginal cost of each traditional shoot compounds into a major line item. Generative 3D content changes the slope of that cost curve, which is why a group of this size can justify a strategic bet on a startup's technology rather than a limited trial.
For a decision-maker, scale is both the opportunity and the discipline. The larger the content operation, the greater the potential saving and the stronger the case for building a repeatable pipeline around digital twins. But scale also means the failure modes matter more. Photorealism has to hold up across a full assortment, brand consistency has to be enforced across thousands of generated assets, and the pipeline has to integrate with existing production and asset-management systems. A pilot that lifts conversions is encouraging; an operating model that governs quality at scale is what actually delivers the return.
The build-versus-partner decision favors specialists here
LPP chose to partner with a focused startup rather than build spatial-AI capability in-house, and that choice is instructive. Photorealistic 3D twin generation is a deep specialty that sits well outside the core competency of a fashion retailer, and the tooling and model expertise required would take significant time and talent to replicate internally. Partnering with Pirxe lets LPP capture the capability quickly while keeping its own teams focused on brand, merchandising, and channel strategy where its advantage actually lies.
The tradeoff a technology leader must weigh is dependency against speed. A strategic partnership with a young startup carries vendor-risk questions around continuity, roadmap control, and how deeply the retailer's creative pipeline becomes entangled with an external platform. Those are manageable concerns, but they should be priced in with contractual and integration safeguards rather than ignored in the enthusiasm of a strong pilot. The general pattern holds: for a capability this specialized and this far from core, buying the expertise usually beats trying to build it from scratch.
What marketing-tech leaders should take from this benchmark
The concrete value of this announcement is that it replaces speculation with a number. Marketing-tech leaders weighing generative 3D product content against traditional production now have a measured reference point: a 21% conversion lift and a 21.6% CTR gain from a real deployment at a major fashion group, plus a company willing to scale on the strength of it. That is a far stronger basis for an internal business case than a vendor's projected ROI, even accounting for the specifics of Sinsay's context.
The roadmap move is to run a scoped pilot that isolates creative-production cost as the variable and measures conversion and CTR against a genuine control, then decide on scale from your own data rather than LPP's. The broader signal is that generative content is crossing from experiment into operating infrastructure at large retailers, and the advantage will accrue to teams that build a governed, repeatable pipeline early. The organizations still treating AI imagery as a one-off campaign trick will find themselves outpaced on both cost and speed of creative iteration.
