Applying artificial intelligence to industrial design requires deciding where it can be most effective and how its results will be evaluated. In Chapter 44 of Toque de Ingenio, Àlex Casabò details his experience incorporating generative tools into the design process and team training. The discussion differentiates between visual exploration, modeling, and manufacturing, and positions the designer’s judgment within the relationship between these phases.
Guest: Alex Casabò, an industrial designer and co-founder of KASAK.
Interview published: April 1, 2025. Episode: 44. Duration: 1 hour and 12 minutes.
In this episode:
- Which design tasks could benefit from the generation and manipulation of images.
- How an experiment with jewelry allows us to understand the process of transforming an image into a physical piece.
- Why building a team requires understanding their work and documenting how they use AI.
Alex describes his encounter with generative tools, drawing on a background dedicated to design, modeling, and rendering. His initial experiments revealed possibilities within images that still had obvious limitations. This article recounts his experience and his assessment of the tools in April 2025, a context that is important to consider when reading any technological references related to this episode.
Industrial design encompasses more than just the final appearance.
The conversation revisits a process that begins with an understanding of the market, the user, and their needs. This is followed by exploration, sketching, modeling, and physical testing. The production and communication of the product then add further decisions to this journey.
This broader perspective helps to situate AI. Generating an image can be useful in an exploratory phase, but it doesn’t, on its own, demonstrate that the object responds to use, that its components fit together, or that it can be manufactured. Each stage requires a suitable method for verifying its results.
Alex recalls this when describing a project carried out using generative tools:
“The design encompasses the entire chain”.
Alex Casabò 34:22.
For a company, this idea allows them to formulate a more useful question than simply choosing a trendy application: what task consumes time, what result the next professional needs, and how to maintain quality. The tool is evaluated based on its contribution to the process. product development.
Explore alternative options and maintain the original selection criteria.
Alex explains that he began by testing different platforms and observing what he could achieve through text instructions. In his early experiments, the results were imperfect, but they demonstrated a new way to explore visual forms and references.
During the episode, we’ll be discussing tools for generating images, transforming sketches, and seeing how a visual concept evolves as it’s drawn. The aim is to speed up certain explorations and make it easier for a concept to be discussed. The image then becomes a working material for the team.
The ability to generate numerous options shifts some of the effort towards selection. It’s necessary to define what’s being sought, recognize what holds interest, and discard results that don’t align with the project’s goals. Simply having a large number of images does not, on its own, establish a design direction.
It’s also important to maintain contact with the references and decisions that led to a proposal. Knowing why a particular form was chosen helps to continue the work when manufacturing, assembly, or usage requirements arrive. Without this continuity, an attractive image can be difficult to turn into a coherent product.
What does a jewelry project created using AI teach us?
Alex describes a project completed within a short timeframe, focused on creating jewelry pieces intended for metal printing. The team gathered references, developed visual options, and selected a set of proposals. Subsequently, they used tools to obtain three-dimensional models, which then needed to be reviewed and refined.
The case demonstrates a complete process: referencing, exploration, selection, model preparation, and fabrication. The designer’s involvement appears in various stages, not just in writing an initial instruction. The final outcome depended on the ability to link the results in a useful and effective manner.
Alex highlights a key aspect of the example: it wasn’t about a product needing to fit precisely with another component through technical geometry. This characteristic facilitated the experiment. The type of object determines how much can be gained from a generated representation and what subsequent work is required.
Teaching a manufacturing team involves evaluating each application based on its specific constraints. Items such as a decorative piece, a casing with joints, and a general application all have different requirements. prototypes They help to determine which aspects of the visual proposal are viable when it is transformed into a physical object.
Team-specific training.
KASAK was created, as Àlex explains, to help research and design departments integrate these tools. Its approach begins with understanding the client and adapting training to their specific needs. The value of a session depends on the type of projects and tasks that the team is undertaking.
In the interview, she discusses the difficulty of keeping courses open on technologies that are constantly evolving. Therefore, she describes a training format that can be updated and adapted to the current context. The priority is to understand how to apply the tools in practice, in addition to learning a specific interface.
This approach is useful for companies with diverse profiles. A person creating commercial images will need something different from someone who models components or defines concepts. Training everyone in the same way may fail to address the specific questions raised by each role.
The change also needs time to test and share results. Introducing a tool into an existing process requires reviewing what information goes in, what comes out, and who will use the resulting information afterward. Training becomes more valuable when it’s linked to a specific project and a decision the team already needs to make.
Learning without losing sight of daily work.
Edgar and Alex are discussing the feeling of being overwhelmed by the constant influx of new information. The conversation acknowledges the pressure to continually discover new apps and to imagine opportunities that aren’t yet being utilized. This feeling can make it difficult to decide where to focus one’s efforts.
Alex introduced a specific element stemming from his own work: at that time, he was still dedicating a significant portion of his work to design-related tasks. The introduction of new capabilities was a gradual process. This experience helped to differentiate the speed of advertising from the speed at which a team transforms its operations.
When speaking about learning, he emphasizes:
“Adaptability is the most important thing”.
Alex Casabò 48:55.
The adaptation described involves testing, reviewing, and continuous learning. To organize this effectively, it’s beneficial to select use cases that make sense for the team and to observe the results. This allows you to accumulate experience without automatically requiring a complete overhaul of the entire process.
Documenting the process allows for a better assessment of the outcome.
The interview concludes by focusing on the training of future designers. Àlex explains that he’s interested in knowing how a student has incorporated AI, and what decisions they’ve made in doing so. However, the final image doesn’t contain all the necessary information to assess the learning process.
Requesting that the process be documented allows us to see references, iterations, and criteria. This practice is also valuable in a company: it makes it easier for another person to understand how a proposal was developed and which aspects still need to be validated.
A visible process helps to use AI with a clear purpose. It allows you to differentiate between exploratory use and a final decision, to identify what the team has reviewed, and to maintain continuity between design and engineering. The result becomes more valuable when it can be explained and further developed.
What can another development team learn?
- Choosing specific tasks where AI can demonstrably improve performance and provide measurable results.
- To differentiate an exploratory image from one that represents a model prepared for production and validation.
- Adapt the training to align with the specific projects and responsibilities of each individual’s role.
- Maintain references, iterations, and criteria so that the process can be reviewed and continued.
The conversation with Àlex Casabò positions artificial intelligence within a craft that connects needs, forms, and production. If your company needs to transform an idea into a product, i-mas can help you structure that journey and choose tests that allow you to progress with sound judgment. Please explain what you would like to develop..
To continue reading: The development process for Citring One..
Source of article: An interview with Edgar Guerrero, hosted by Àlex Casabò on Toque de Ingenio, published on April 1st, 2025. The quotes link to the relevant sections of the video; the “learnings” are a summarized editorial version of the conversation.
