Digital transformation through systems thinking

Layered abstract waves in turquoise, coral and pale green

A fashion business can introduce a new piece of software and still leave the most expensive part of its process untouched. That is why, in my work on digital transformation, I tend to begin with the connections between activities rather than the tool a team wants to buy. A change in design can affect the work of a supplier; a more useful product record can change what happens in a warehouse, in a store and after the sale. The value depends on how those changes fit together.

My experiences at Versace and Giorgio Armani gave me different ways to work on those connections. They also made me wary of describing transformation as a sequence of technology launches, because a technically successful implementation can still leave people with duplicated work or a process that no one has the time to use.

Transformation reaches beyond the familiar digital departments

Marketing and e-commerce were among the first functions expected to deliver digitally, but the pandemic made the need much harder to ignore in wholesale, merchandising and operations. A virtual showroom, for example, changed more than the format of a buyer meeting: it depended on product information, the availability of images or samples, and the way a collection was presented.

I have explored this broader responsibility before in The democratization of digital delivery. Business functions understand many of the problems they need to solve, yet the people doing the work can be too occupied with daily operations to investigate partners, gather requirements and manage a new implementation. An innovation and digital transformation function can help turn that practical knowledge into a workable project, in collaboration with the teams whose routines will change.

For me, digital transformation means changing how the organisation works through the use of digital technologies. Its possible benefits include less manual work, fewer errors and more useful information for decisions, but none of those follows automatically from installing software. The work also involves responsibility for the process and the capacity to support people while it changes.

Diagram showing technology, culture and processes as connected dimensions of organisational change
Technology, culture and processes are connected dimensions of organisational change. View original diagram.

Systems thinking makes the dependencies visible

Systems thinking means looking at how the parts of a system influence one another, including its information flows, feedback and constraints. In a business, this involves tracing what happens before and after a task and noticing where a decision in one department creates a problem somewhere else.

Donella Meadows, an environmental scientist and systems thinker, described information flows, rules and goals as places where interventions can change the behaviour of a system. Her work helps explain why improving the handover between teams may matter more than making one isolated task faster. [1] I use that lens here to read my own project experiences; I am not suggesting that the teams involved followed a formal Meadows methodology.

The practical consequence is to choose an appropriate scope for the work. A production improvement needs to account for what design requires and what the supplier can deliver. A customer-facing service needs reliable product information and an operating process behind it. Optimising one of those pieces while ignoring the others can simply move the inconvenience to another team, where it will appear on a different budget line and acquire a new owner.

Versace: changing the sequence of printing and cutting

At Versace, I worked on a project involving the placement of prints on fabric. The existing process used paper patterns and pins to position the shapes needed for a garment. Once the printed fabric was cut, the material outside those shapes also contained ink that had never needed to be there.

We introduced software to place the print and the required shapes digitally, so that printing could focus on the areas that would become part of the garment. We then looked at the unused space between those shapes, because better placement could also reduce the area of fabric needed. The project connected the decisions made in design with printing and cutting, rather than treating fabric waste as a problem to address only at the end of production.

The intention was to give the people doing skilled work better tools while retaining the material, ink and quality requirements. I still find the example useful because it shows how the scope of a digital intervention changes when the sequence of work is reconsidered.

Versace: 3D beyond the final image

At Versace, we used 3D during lockdown, when closed suppliers could prevent the production of physical samples and buyers could not travel to a showroom. In one example, a printed detail on a pair of trousers was absent from the physical prototype and its photograph, but could be represented on the 3D garment after a late merchandising and design decision.

That was a concrete use of 3D to deal with a missing piece of the sales presentation. The broader opportunity was to bring the digital asset into the earlier stages of work, connecting the idea, prototype and collection sample to the image used later. A model created upstream could have uses in more than one department, instead of each department commissioning another representation of the same product.

There was also a limit: designers may need to touch or examine a physical product before making a decision. A model can support that decision without meeting every requirement that a physical sample serves. These virtual workflows opened up possibilities, but did not mean that the company had already replaced its end-to-end production process.

Armani: the photoshooting process starts before the studio

The photoshooting workflow is another Armani example, which I explored in Digital asset production in luxury fashion: the AI-led revolution. The useful starting point is the journey of a product and its information, rather than the moment a photographer presses the shutter. Sample availability and buying decisions affect the shooting schedule; studio capacity and the readiness of a sample affect what can be captured; approval and post-production affect when the resulting image can reach e-commerce or a retail partner. [3]

Following those dependencies exposes work that disappears in a discussion focused only on photography. A sample that cannot be located creates a scheduling problem. A photographer handling avoidable set logistics has less time for creative work. An approved image that travels separately from the product data creates another round of manual tagging before it can be published. Making one stage quicker can leave the asset waiting at the next handover.

The relationship between wholesale and consumer-facing content matters too. B2B and B2C teams can commission representations of the same product under different guidelines, so the opportunity is to establish where an existing asset is genuinely suitable for reuse and where another image is necessary. Reuse without checking creative and product requirements would create a different kind of rework.

Connecting a Product Information Management system, or PIM, with a Digital Asset Management system, or DAM, addresses the relationship between product information and its images. AI-assisted tagging, retouching and other studio automation may help with parts of that work, but the requirement remains an accurate representation of the garment with appropriate human approval. The technologies discussed in the photoshooting article are opportunities to evaluate, not a claim that Armani implemented every automation described there. This is why I read the example as a process-design problem before treating it as an AI procurement decision.

Armani: traceability, the Digital Factory and configurators

My later work at Giorgio Armani, described in my professional experience, included RFID and digital product passports, 3D prototyping, virtual showrooms and online configurators. They followed my earlier work at Versace. [2]

With RFID integrated into product labels, traceability could be considered across several functions. Production and logistics needed information about a product’s progress, while retail needed a usable stock-taking process. A QR-linked product record could also give customers access to information about the garment, including its sustainability story. Those uses depended on the quality and ownership of the underlying data; a code on a label would have been a poor substitute for either.

The Digital Factory work brought another set of relationships into view. Virtual prototypes could support design review and changes before a physical sample was available, while 3D assets could have further uses in a showroom, in marketing imagery and in e-commerce configurators. The benefit of reusing an asset depended on whether it met the requirements of those different functions. A prototype suitable for an internal discussion is not automatically ready for a customer-facing product page.

The online configurator for made-to-measure services addressed a related service problem: allowing measurements and product choices to support repeat orders, rather than requiring the customer to begin the fitting process again each time. The intended convenience had to be matched by the accuracy of the information and the handover to the people making the garment. When a service retains a customer’s measurements, the way that information is handled also deserves the same attention as the interface.

Across these projects, I was interested in making information useful at the point where someone needed to act on it. Faster stock-taking, a more usable prototype and a smoother ordering process are reasonable objectives, but I do not want to turn them into universal promises or repeat percentages without the scope and evidence needed to explain them.

A technically interesting idea can still be the wrong project

Another project I tried to advance involved AI-based quality control using computer vision and machine learning. The investment required even for a proof of concept was too high for the project to progress as I had hoped, and it remained a failed attempt, not a deployed solution.

That belongs in the account alongside the projects that moved forward. A plausible technical solution does not settle the question of affordability or timing, just as a digital prototype does not remove the designer’s need to handle a fabric. For me, taking those limits seriously is part of transformation work, because the organisation still has to operate while an idea is being tested.

The approach I bring to digital strategy and transformation starts with the work and follows its dependencies. I want the people using a new process to understand what it changes, the teams supporting it to have clear responsibilities, and the business to be able to judge the result against an objective it actually cares about. Systems thinking gives me a way to connect those decisions without pretending that one more platform will make them for us.

Updated on 12 November 2026. The original publication date is retained.

References

  1. Donella Meadows, Leverage points: places to intervene in a system
  2. Fabrizio Rinaldi’s current professional experience and CV
  3. Fabrizio Rinaldi, Digital asset production in luxury fashion: the AI-led revolution, 13 February 2025

Discover more from INNOVATION WAVES

Subscribe now to keep reading and get access to the full archive.

Continue reading