The intermediate results from the first reporting period already show progress beyond the state of the art in several core areas of demand-side flexibility management. First, the project has moved beyond isolated algorithms and conceptual architectures by delivering and validating an initial IDOP that combines Customer Energy Management Systems, Aggregator Energy Management Systems and deployment support into a coherent, standards-aligned framework. This advances beyond current fragmented practice, where interoperability across legacy building systems, customer-side optimisation and aggregator-side market participation is typically handled through bespoke integrations. Second, in resource modelling and control, the project has advanced beyond conventional grey-box approaches by developing and testing a Neural ODE Toolbox with hybrid modelling methods, including latent and physics-informed Neural ODEs, and applying them in real pilot contexts such as residential heat pumps, a dairy farm and a supermarket HVAC system. These results improve modelling accuracy and robustness while creating a practical pathway towards data-efficient optimisation and control of flexible assets under real operational constraints. Third, the project has progressed beyond current interoperability practice by defining and implementing standard-oriented interfaces on both the customer and aggregator sides, while also developing automated, LLM-supported approaches for semantic adaptation of legacy systems. In parallel, the AEMS developments extend beyond single-market approaches by providing initial probabilistic forecasting and multi-market bidding capabilities, integrated with operational platforms and validated against real market data.
At this stage, these advances should be seen as concrete intermediate results that reduce key barriers to large-scale uptake of demand-side flexibility management. The pilots in four countries show that the approach can already be deployed in representative residential, commercial and industrial environments and validated at TRL 7, while providing early evidence of scalability, standards compliance and multi-market applicability. The main priorities for the next phase are broader field qualification, stronger standardisation alignment, smoother integration into existing market and operational processes, and continued technical refinement. In particular, the project will need to consolidate phase 1 results into fully qualified TRL 8 solutions, expand semantic adapter automation across more legacy systems, further validate forecasting and optimisation across asset classes and markets, and continue interaction with standardisation and market actors to ensure replicability beyond the pilots. If these next steps are achieved, the project results have strong potential to accelerate the deployment of interoperable, secure and economically viable flexibility management solutions for buildings and industrial sites across Europe.
