Digitise product development to cut sampling waste

Submitted through the COP31 Sustainable Transformation Awards · SKD Türkiye (WBCSD Global Network Partner)

The company is an apparel retailer operating in 100 countries through more than 500 stores and digital sales channels, with a workforce of more than 1,000 people.

Fashion retail carries a resource-intensive product development chain. A style is normally proved through physical samples that have to be cut, sewn, freighted and reviewed, and every product image is normally produced through a studio shoot with models, sets, lighting and the movement of garments and crew. Both stages consume fabric, energy, transport and time before a single garment is sold, and both are repeated for every collection.

The company treated this as a resource question rather than a creative one. The objective was to compress the route from concept to technical drawing, from virtual fitting to marketing image and on to publication on the e-commerce channel, while reducing the fabric, studio energy and freight consumed along the way.

The initiative sits inside the company's 5.0 corporate vision and its Triple-D strategy, under which sustainability is handled as a strategic transformation heading that shapes decision-making, the objectives and key results system, managers' target cards and business unit priorities. It has passed the pilot stage and has become an established way of working in design, product development, sample approval and e-commerce content production.

Environmental performance is tracked under the GHG Protocol across Scope 1, Scope 2 and Scope 3 emissions, and selected environmental and social indicators are subject to independent limited assurance under the ISAE 3000 standard.

Location of the initiative: Türkiye, applied across an operation covering 100 countries and more than 500 stores

The ecosystem has two connected parts, both built on artificial intelligence and three-dimensional design, and both aimed at the stages of product development that consume the most material and energy.

The first part is digital design. Designers generate photorealistic prototypes from text-based prompts and test alternative fabrics, patterns and forms against each other before anything is cut. Virtual fitting allows the fit of a style to be assessed before production, which turns the decision to make a physical sample into a deliberate choice rather than an automatic step.

The second part is digital content production. Product imagery is generated with artificial intelligence supported models instead of a physical shoot, which removes the model agency booking, the studio hire, the set build, the lighting load and the movement of garments to and from the studio. New collections can be made ready for sale in days rather than weeks.

The two parts are joined into a single chain. A style travels from concept to technical drawing, virtual fitting, marketing image and e-commerce publication without leaving the digital environment, so the data created at the design stage is reused rather than rebuilt at each handover.

Supplier alignment extends the same chain outside the company. Pattern development and three-dimensional integration work is carried out with commercial three-dimensional design and pattern technology providers, and pattern makers at supplier factories work from the same technical files, measurement charts and pattern data, so that a fit approved digitally holds in the factory.

Representation is designed into the image side rather than added afterwards. Artificial intelligence models representing different body types and ethnic backgrounds are used in visual production, and three-dimensional fit work is carried out for plus sizes.

The distinguishing feature of the model is where the sustainability effect is created. It is not sought in the product material alone but at every stage of the value chain, by optimising resource use through digital tools at the point where the resource would otherwise be consumed.

Figure 1: Flat technical sketch of the style

Figure 2: AI generated product visual of the same style

Sustainability Impact

Climate

Environmental performance is tracked under the GHG Protocol across Scope 1, Scope 2 and Scope 3 emissions. The initiative acts on more than one of these at once: studio lighting and equipment energy sits inside the company's own operational footprint (Scope 2), while sample production and bought-in production services fall under Scope 3 Category 1 (purchased goods and services) and the transport of garments and samples, together with other upstream logistics, falls under Scope 3 Category 4 (upstream transportation and distribution).

The measured lever is avoided physical sampling. In 2025, 667 styles were fitted in a three-dimensional environment, which prevented 1,334 additional samples from being produced and shipped, together with the fabric, cutting, sewing and freight each of them would have required.

Digital content production removes a second energy load. Replacing studio shoots with artificial intelligence generated imagery removes studio energy, production set-up and the logistics of moving garments, models and crew, and produced a saving of between 70 and 90 per cent in model agency, studio and production costs.

Impact data for 2025 is monitored within the company's reporting period through digitalisation investment, research and development and Design Centre investment, software development projects, three-dimensional modelling gains and cost saving data from the ecosystem, in parallel with base-year emission targets. Selected environmental and social indicators are independently assured under the ISAE 3000 standard, and the company reports to CDP on climate, water and forests.

Nature

Preventing 1,334 additional samples avoids the fabric consumed in making them, which is a material saving upstream of the company rather than a waste stream managed after the event.

Because alternative fabrics, patterns and forms are compared in a digital environment, the iteration that would normally consume cloth is carried out without cloth. Resource efficiency is therefore created at the design decision rather than at the disposal stage.

Reduced sample movement and the removal of studio shoots also remove the freight legs associated with them, which lowers the transport load attached to each collection.

Social

Digital capability inside the workforce has increased, because designers, product developers and content teams now work with three-dimensional and artificial intelligence tools as a routine part of the job rather than as a specialist add-on.

Visual production has become more inclusive. Artificial intelligence models representing different body types and ethnic backgrounds are used in product imagery, and three-dimensional fit work is carried out for plus sizes, so the range of customers represented and fitted widens rather than narrows.

Capability is also transferred outside the company. In 2025, 83 supplier firms received a total of 159 hours of training across 17 different categories. Seminars held with students from Marmara University raised awareness of three-dimensional design technologies among young talent, and collaborations with Haliç University and Dokuz Eylül University covered brand, design, patent and copyright processes.

Business Impact

Benefits

The main commercial benefit is speed. New collections can be made ready for sale in days rather than weeks, because the image production step no longer waits for a studio, a model booking and the physical garment.

Cost falls with it. Digital content production delivered a saving of between 70 and 90 per cent in model agency, studio and production costs, and three-dimensional modelling produced a gain of more than TRY 2.3 million.

Avoided sampling removes cost that never appears on an invoice as a saving: the 1,334 samples prevented in 2025 carried fabric, labour and freight that were simply not spent.

Completion of the full programme of work is expected to deliver a cost saving of approximately TRY 25 million. This figure is a projection tied to completion, not a realised result.

Operational agility improves in a less visible way. Because design, fit and content data live in one chain, a change made at the design stage propagates to the technical file and the marketing image instead of triggering a new round of physical work.

Costs

The cost base is investment rather than operating expense. Digitalisation investment stands at approximately TRY 377 million, research and development and Design Centre investment is above TRY 201 million, and the software development portfolio comprises 407 projects.

Operating cost shifts rather than disappears. Studio, model and freight spend falls, while software licensing, computing capacity, data management and the internal capability to run artificial intelligence and three-dimensional tools rise.

The model carries three dependencies that behave as costs if unmanaged: the accuracy of artificial intelligence outputs, the speed at which teams adapt to new tools, and the digital maturity of suppliers. These are handled through training, internal capability development, controlled rollout and a phased coordination model with suppliers.

Costs are contained by building the solutions inside the group's own technology unit, which develops software, data analytics, artificial intelligence, cloud, business intelligence and e-commerce infrastructure, and by reusing the same digital asset for fit approval, marketing imagery and e-commerce publication instead of producing each separately.

Impact Beyond Sustainability And Business

Co-benefits

The supplier training effort turns an internal capability into a sector capability. Strengthening technical files, measurement charts and pattern compatibility with supplier pattern makers raises production accuracy for the supplier's other customers as well.

University collaborations extend the same effect to the labour market. Seminars with Marmara University students and the work with Haliç University and Dokuz Eylül University on brand, design, patent and copyright processes build three-dimensional design skills before graduates enter the sector.

Inclusive imagery has a reach beyond the company's own catalogue, because representation of different body types and ethnic backgrounds in mainstream apparel content changes what customers see as normal.

Potential side-effects

Accuracy of artificial intelligence output is the principal risk. A prototype or a product image that misrepresents the finished garment transfers a quality problem to the customer, so a human review gate has to sit between generation and publication.

Virtual fitting reduces physical sampling but does not remove it. New fabrics, unfamiliar constructions and categories where digital simulation is less reliable still require a physical sample, and treating the digital result as final in those cases would raise return rates rather than lower them.

Supplier digital maturity varies. Suppliers that cannot work from digital technical files gain nothing from the chain and can become a bottleneck, which is why the rollout is phased rather than uniform.

The skills mix changes. Pattern and content work moves towards digital tooling, which requires retraining rather than replacement if the craft knowledge held by experienced staff is to be retained.

Typical Business Profile

The model suits apparel and fashion retailers that carry an in-house design and product development function, a formal sample approval process and an e-commerce channel that consumes large volumes of product imagery.

It is most relevant where collections are frequent and assortments are wide, because the saving is created per style and therefore scales with the number of styles developed each season.

Delivery engages the design centre, the technology unit, product development, e-commerce, marketing, supply chain and sustainability functions working to a shared plan, and requires a supplier base able to work from digital technical files.

Approach

1. Map the resource load of product development: Record how many physical samples, studio shoots, model bookings and freight movements each collection consumes, so the stages carrying the highest fabric, energy and logistics cost are known before any tool is bought.

2. Build one digital chain rather than isolated tools: Connect text-based prototype generation, technical drawing, virtual fitting, marketing imagery and e-commerce publication so a style moves through a single environment instead of being handed between disconnected systems.

3. Set a rule for when a physical sample is still needed: Use virtual fitting to assess fit before production and treat the physical sample as an exception that has to be justified, because the avoided material use comes from the decision rule and not from the software.

4. Replace studio shoots selectively: Start with the product categories where artificial intelligence generated imagery reaches the required realism and keep physical shooting where it does not, so image quality is not traded for speed.

5. Design representation into the image library: Build artificial intelligence models covering different body types and ethnic backgrounds and run three-dimensional fit work for plus sizes, so digital content widens rather than narrows the range of customers shown.

6. Align suppliers on the same digital data: Work with supplier pattern makers on technical files, measurement charts and pattern compatibility so a fit approved digitally holds in the factory, and phase the rollout to match each supplier's digital maturity.

7. Train people, not only systems: Deliver structured training internally and to suppliers, at the scale of 159 hours across 17 different categories to 83 supplier firms in 2025, and run university seminars so the skills base extends beyond the company.

8. Measure avoided volume, not only cost: Track styles fitted digitally, samples prevented, production cost saving and time to market, report the results inside the corporate sustainability reporting cycle and place selected indicators under independent limited assurance.

Stakeholders Involved

• Project leads: Ownership sits at the most senior level through the company's 5.0 vision and Triple-D strategy, under which sustainability is treated as a strategic transformation heading that shapes decision-making, the objectives and key results system, managers' target cards and business unit priorities. Corporate integration is carried by the technology unit, the Design Centre and the leadership responsible for sourcing, sustainability and corporate social responsibility, working together rather than through a single owner.

• Company functions: The Design Centre, the technology unit, product development, e-commerce, marketing, supply chain and sustainability teams share the rollout. The technology unit supplies the technical continuity of the ecosystem through software development, data analytics, artificial intelligence, cloud technology, business intelligence and e-commerce infrastructure, while product and content teams define the requirements the tools have to meet.

• Main providers: Commercial providers of three-dimensional design and pattern development technology supply the digital infrastructure and work with the company on pattern development, three-dimensional integration and digital alignment with suppliers.

• Other: Supplier factories participate through their pattern makers, and 83 supplier firms received 159 hours of training across 17 different categories in 2025. Marmara University students took part in seminars on three-dimensional design technologies, and collaborations with Haliç University and Dokuz Eylül University covered brand, design, patent and copyright processes. Independent assurance providers verify selected environmental and social indicators under the ISAE 3000 standard.

Key Parameters To Consider

The initiative has passed pilot stage and now functions as an institutionalised working model in design, product development, sample approval and e-commerce content production.

Continuity is secured financially and managerially by digitalisation investment close to TRY 377 million, research and development and Design Centre investment above TRY 201 million, and a portfolio of 407 software development projects.

The rollout is staged. In the short term the model extends to further categories including plus sizes; in the medium term technical files, measurement charts and pattern compatibility are strengthened with supplier pattern makers; in the long term the intention is a systemic working standard across the fashion value chain.

The main risks are the accuracy of artificial intelligence outputs, the adaptation of teams to new tools and the varying digital maturity of suppliers. These are managed through training, internal capability development, controlled scaling and phased supplier coordination.

Implementation And Operations Tips

Counting avoided samples is what makes the environmental case visible. Cost saving alone understates the result because it does not capture the fabric, the cutting and sewing labour or the freight that were never committed.

The saving lives in the decision rule rather than in the software. Unless the organisation agrees that a physical sample now needs a justification, the digital tools sit alongside the old process instead of replacing part of it.

A human review gate between generation and publication is not optional. Artificial intelligence output that misrepresents a garment converts a design saving into a customer return.

Supplier alignment should be sequenced by digital maturity. Pushing digital technical files onto suppliers that cannot use them produces rework and erodes trust in the model.

External Links

• Sustainability report

• SBTi validated targets

• CDP disclosure

• Independent third-party external assurance