Can we use AI to accelerate the development of low-carbon construction materials? The Reincarnate demonstration on circular feedstock explores this possibility. Led by Ragn-Sells, the case focuses on the use of the SLAMD platform to design mortar formulations that replace a significant share of cement with recycled glass and concrete fines.
Download the one-pager
At the core of the demonstration is a digital laboratory environment combined with sequential learning algorithms. Instead of relying on traditional trial-and-error, SLAMD iteratively proposes new material recipes, which are then tested and fed back into the system. Over six optimisation cycles, 32 different formulations were developed and evaluated under real laboratory conditions, targeting a minimum of 30% cement substitution while maintaining a compressive strength of at least 30 MPa.
Download the three-pager
The results show that the approach can effectively guide material development and improve predictive accuracy over time. While early cycles were affected by variability in laboratory conditions, the system stabilised in later stages, reducing prediction errors and identifying viable low-cement mixes. By the final cycle, the model outperformed its initial baseline, demonstrating its capacity to learn from limited experimental data and narrow the range of outcomes.
Beyond the individual formulations, the demonstration highlights the potential of AI-driven tools to significantly shorten development time and enable new circular value chains based on secondary raw materials. By integrating digital modelling with laboratory testing, SLAMD offers a scalable approach to designing more sustainable binders for the construction sector.