AI is revolutionizing the sustainable building materials industry, but how exactly does it work?

Fear not! This text decodes the magic behind SLAMD, our data-driven material design tool, and makes its revolutionary power accessible to all.


Mastering material recipes, SLAMD is your ultimate AI-Chef!

In the world of cooking and materials development, the creation of a meal or a material is essentially the development of a recipe. Just as a chef combines specific ingredients in precise proportions to create a delicious dish, traditional construction materials like concrete are developed using a similar methodology. The construction industry is now facing immense challenges, and these classic recipes no longer meet the growing demands for sustainability, resource-efficiency, and cost-effectiveness. Better “meals” are possible, but they have often been developed by chance. Optimizing recipes can be challenging, as many constituents have a relatively uncertain effect on the end result. This is especially true for complex creations. 


The Complexity of sustainable Recipes

The difficulty in developing innovative recipes arises from the countless variations and the exponential correlation between the number of ingredients and the possible results. Similarly, in the construction sector, finding a suitable alternative to environmentally harmful cement involves exploring a wide range of options, such as natural binders or industrial byproducts. Finding the optimal blend without sacrificing performance or sustainability is a major challenge.

At first glance, one might assume that inventing a new pizza is as simple as changing the toppings – et voilà! However, this assumption misses the point. The real challenge lies in modifying the basic composition of the most important pizza component: the dough. How, for example, could one develop a dough that would have an extended shelf life of three days? In materials science, the discovery of new properties is often based on chance and trial and error.


Data driven Materials design

Data-driven material design revolutionizes the process by creating a detailed and comprehensive digital “cookbook” of all feasible recipes. At first glance, this approach may seem counterproductive, as it appears excessive to compile an entire library of cookbooks for optimizing a single dish. However, for AI-driven material design, this method is a game-changer. Rather than attempting to predict the perfect recipe — an insurmountable challenge — the task is transformed into a much simpler search problem, seeking the recipe that best aligns with the material requirements. Non-obvious solutions are naturally integrated, a phenomenon that could be considered systematized luck. Previously, data-driven approaches faced difficulties in finding practical applications due to the large amount of training data required. However, the “cookbook trick” employs highly efficient AI algorithms that necessitate only a minimal amount of training data, making the approach far more viable.

The “Cookbook´´ trick: Surprisingly effective, it conjures thousands of recipes to perfect just one meal!

The Cookbook Trick and the Digital Lab Workflow

SLAMD’s Digital Lab ingeniously exploits complex constituents and processing data information. By scaling up this data, a comprehensive cookbook of possible recipes is created (so called search space or input space). The options are evaluated based on their potential utility, while also considering the sustainability and cost implications of each constituent and process. This evaluation enables the discovery of a balanced mix of sustainable and efficient material solutions.


The AI Chef

SLAMD’s AI optimization is like a master chef, skillfully matching requirements and pr

edicted properties to develop tailored, sustainable, and cost-effective building materials. The design process focuses on precise requirement specifications and development strategies. Formulations can be iteratively adjusted based on the latest insights from the material lab. With a keen eye for detail, SLAMD prioritizes materials that are both environmentally friendly and wallet-friendly, making it a powerful and adaptable approach to developing innovative, breakthrough solutions.

SLAMD’s landing page provides a user-friendly guide through the AI-driven material design workflow, showing the digital lab on the left and AI optimization on the right. Experience it for yourself with our web preview: 



Reinventing products such as pizza or concrete is an extremely challenging task. Data-driven design frameworks like SLAMD offer a powerful approach to revolutionizing classics by providing an adaptable and tailored method for developing sustainable, resource-efficient, and cost-effective building materials. By leveraging the digital lab workflow (digital cookbook approach), and AI-Optimization, SLAMD paves the way for a more sustainable, adaptable and innovative future in construction.


Main author: Christoph Völker, Data Scientist Researcher at BAM

Join the Reincarnate Community of Practice by subscribing to our Newsletter, follow us on social media — LinkedIn and Twitter —, and connect with us to explore collaboration opportunities.