Figure 1 · Generative Product Promotion in Operation The bespoke promotional tool in operation, applying generative AI to place an additive-manufacture footwear concept directly onto a human model and generating a full commercial video advertisement.

Creative Computation to Commerce

Generative design, making, and market: a postgraduate module restructured as a design-to-enterprise pipeline

Computation to Market Level 7 Postgraduate Additive Manufacturing & AI 6 Trading Ventures 14 Commissions

Overview

Computational design teaching usually ends at the same point. The student generates a form. The form is rendered, printed, and presented. The module concludes.

This is a curious place to stop. The student has produced an artefact with genuine market potential. They have no idea how to test that potential, and the curriculum has not asked them to. Value creation is treated as somebody else's subject.

The module was restructured to remove that boundary. Generative design, physical making, and commercial launch were treated as one continuous process rather than three separate concerns. Students generate parametric forms using AI-powered generative algorithms. They prototype those forms through 3D-printing. They then take selected outputs to market.

The final stage presented an obstacle. Postgraduate design students rarely have access to professional product photography, models, or advertising budgets. A prototype photographed on a desk does not read as a product. The gap between a good object and a saleable one is largely a gap in representation.

I therefore developed a generative-AI tool to close it. The tool places a product on a human model and produces both a commercial still image and a video advertisement. The barrier to market entry fell from a professional photoshoot to a single upload.

The pipeline is the argument. Students do not learn generative design and then, separately, learn about enterprise. They discover that a design decision made in an algorithm has consequences for what can be printed, what can be worn, and what will sell.

The Pipeline

The module runs as three connected stages. Each stage constrains the one before it:

  • Stage 1: Generate. Students develop parametric definitions and apply AI-powered generative algorithms to produce families of forms rather than single objects. The emphasis falls on variation and selection. A student who produces one ring has made a decision. A student who produces two hundred rings must learn to defend one.
  • Stage 2: Make. Selected forms are prototyped through additive manufacture. Making is where computational ambition meets material reality. Wall thicknesses fail. Overhangs collapse. A geometry that resolves elegantly on screen may be unwearable in the hand. Students iterate between definition and print until the object survives contact with a body.
  • Stage 3: Market. Prototypes are then positioned as products. Students develop a proposition, a price, and a visual identity. Selected outputs are advertised using the promotion tool described below, and, in a number of cases, sold through retail platforms that the students established and continue to run.

The Tool: Generative Product Promotion

The tool takes two inputs. The first is an image of the product, whether a piece of jewellery, a garment, or a shoe. The second is an image of a human model. It returns a commercial photograph of the product worn by that model, followed by a short video advertisement in which the model presents it.

Interactive Tool Demonstration · Generative Product Promotion Pipeline The promotional tool synthesising worn commercial visuals and high-definition video assets directly from an additive-manufacture footwear prototype and model input.

Two properties matter pedagogically:

  • Speed. A student can test a product against several presentations in an afternoon. Marketing becomes an iterative design activity rather than a final and irreversible task.
  • Honesty about limits. The tool produces a persuasive image. It cannot produce a proposition, a price point, or an audience. Students discover quickly that a beautiful advertisement for an ill-conceived product simply fails more attractively. That discovery is the point of the exercise.

Pedagogic Structure

The module proceeds in four assessed stages:

  • One: Computational fluency. Students acquire parametric and generative methods through short technical exercises. Work is their own from the outset. No student proceeds with a definition they cannot explain.
  • Two: Form-finding and selection. Generative algorithms produce large solution spaces. Students must articulate the criteria by which they narrow those spaces. Selection is assessed as a designerly act, not a matter of preference.
  • Three: Prototyping and iteration. Physical printing tests the geometry. Failures are documented and treated as evidence. Students return to the definition, revise it, and print again. This loop is the experiential core of the module.
  • Four: Proposition and market. Students define an audience, a price, and a route to sale. Promotional material is produced using the tool. Students who wish to trade are supported in establishing a retail platform, and several have done so.

The sequence is deliberate and cannot be reordered. A student cannot market an object they cannot make. A student cannot make an object they cannot generate. Commercial thinking arrives last because it depends on everything before it.

Learning Theory in Practice

The structure aligns with established learning theories across four dimensions:

  • Experiential learning. The structure follows Kolb's experiential cycle closely. Students act, observe the consequences, form a revised understanding, and act again. The 3D printer is an unusually honest source of concrete experience. It does not accept an argument, and it does not award partial credit.
  • Constructivism. The approach is constructivist in its treatment of knowledge. Students are not given a market. They construct one, through decisions about audience, price, and presentation that they must then defend. Meaning is made by the learner, and tested against a public.
  • Connectivism. It is connectivist in its treatment of capability. The competent designer here operates within a network of algorithms, machines, generative tools, and customers. Knowing what to ask of each is itself the skill. Students learn to distribute their thinking across that network without surrendering judgement to it.
  • Entrepreneurial learning. The module carries an entrepreneurial dimension that these theories accommodate but do not fully describe. Students are asked to act under genuine uncertainty. Nobody, including me, can tell them whether their product will sell. That condition cannot be simulated in a seminar. It can only be entered.

Outcome & Impact

Three cohorts have completed the module to date with exceptional outcomes:

  • Six student-led retail ventures established and traded. These span jewellery, clothing, and footwear.
  • Fourteen students secured further design commissions after graduation, on the strength of work developed within the module.
  • Over 90% positive feedback in student evaluation consistently across cohorts.

The venture figure is the more demanding of the measures. Satisfaction records how a module was experienced. A trading venture records a graduate acting on their own account, in public, with something to lose. It is an employability outcome that does not depend on an employer.

3D-printed necklace produced parametrically in module
Figure 1a · The Physical ProductA 3D-printed necklace produced within the module. The geometry is generated parametrically and resolved as an articulated chain of linked forms.
Model photograph input provided to generative promo tool
Figure 1b · Model InputThe second input supplied to the promotion tool. Neither image contains the other; the composite is produced entirely by the tool.
Generated commercial photograph showing necklace worn by model
Figure 1c · Generated AdvertisementThe necklace placed on the model, lit and framed as a commercial photograph. Scale, drape, and shadow are inferred automatically.
Figure 1d · Commercial Video Campaign (Necklace) Automated video advertisement synthesised by the promotional tool, animating the human model presenting the articulated 3D-printed necklace in motion.
Parametric open-lattice ring designed for additive manufacture
Figure 2a · The Parametric RingA second product, generated as a member of a larger solution family. The open lattice places material only where structurally necessary.
Model hand photograph selected for ring presentation
Figure 2b · Model InputThe corresponding model hand image. The hand position was selected by the student to match the ring's delicate scale.
Generated advertisement of parametric ring worn on finger
Figure 2c · Generated AdvertisementThe ring rendered in worn context. The two product examples share no visual language or styling, showing the tool imposes no house aesthetic.
Figure 2d · Commercial Video Campaign (Parametric Ring) Dynamic video advertisement produced for the parametric ring, demonstrating specular light interaction and photorealistic hand movement.
Significance

Computational design should not end at the print bed

The approach rests on a single argument: a postgraduate designer who can generate and fabricate a product should not be prevented from selling it by the cost of a photograph, and a curriculum can be built so that they are not. Five claims follow from the work.

  1. Curriculum Innovation Crossing Disciplinary Boundaries

    Computational design, digital fabrication, and enterprise are ordinarily taught by different people in different modules, if they are taught together at all. Here they form one sequence, in which each stage is disciplined by the next.

  2. Tool Development Preceding Pedagogic Need

    The commercial stage of the pipeline was not possible with available software at acceptable cost. I identified the obstacle and built the tool that removed it. The pedagogic requirement preceded the technology and determined its form.

  3. Entrepreneurial Learning with Measurable Results

    Six ventures were established and traded. Fourteen graduates secured further commissions. These are not simulated business plans marked against a rubric. They are verified outcomes in a competitive public market.

  4. Experimentation & Iterative Development as Assessed Practice

    Students work through large generative solution spaces, fail repeatedly at the print bed, and test propositions against an audience. Iteration is structured directly into the assessment and evidenced in the work.

  5. Habits for Complex Real-World Challenges

    Graduates leave with capability in emerging technologies and in human-centred design, and with the mindset to move an idea from algorithm to object to market. That combination is what the pipeline exists to produce.