Professor Dr. Schlicht, many buildings already have digital systems and data. Why does this still too rarely result in real added value in practice?

This is one of the central paradoxes of our time in the building sector: we have invested heavily in sensor technology, building automation and CAFM systems in recent years and yet the operational added value often fails to materialize. The reason is not a lack of data, but its structural isolation. The historically evolved architectures in existing buildings are primarily designed for local optimization. A heating controller functions reliably within its system, but it does not “talk” to the energy data portal, not to the property management system and certainly not to the user context. What we see in the field is a large number of highly efficient stand-alone solutions that do not allow for overall optimization.

There is also a tangible data problem: although a lot of data is available, it is hardly usable in unprocessed form. A property manager requests monthly consumption data from the FM service provider by email, sorts tables and follows up information with phone calls, a process that experts in the real estate industry have long described as outdated. These manual processes do not generate a reliable database for AI applications that require a large, qualitatively validated data stream.

In short: data and digital systems are available, but the ‘glue’, i.e. a semantically enriched, interoperable layer that bridges silos, is almost completely missing in the existing system. This is precisely where the federated data ecosystem concept, as pursued by SmartLivingNEXT, comes in.

Sources: SmartLivingNEXT (2025): Flagship project quarterly report Q4/2025; Facility Management trade journal (2023): “Making digital building data more accessible”; Memoori (2025): “IoT Platforms in Smart Commercial Buildings 2025-2030”

You are working intensively on the transformation of the portfolio. What role does the ability to merge data from different sources play in this?

In my opinion, the ability to merge data across sectors is the key competence par excellence, both technically and organizationally. Let’s take a simple example: a predictive maintenance system for an elevator system in a residential building needs at least three data sources at the same time: IoT sensor data (vibrations, travel hours), the ERP system of the FM service provider (maintenance history, contract conditions) and, ideally, user feedback from a tenant app. Only the combination of this data results in a robust forecasting model.

Current studies provide impressive evidence of what is possible: predictive maintenance implementations show payback periods of one to three years in practice, with maintenance cost reductions of around 15 percent in smaller residential buildings and energy savings of up to 25 percent with optimized HVAC systems in larger properties. Annual savings of 60,000 to 100,000 euros are documented for an office complex with 50,000 m².

But the trick is that these results can only be achieved if data is allowed to flow across domains. The challenge is not primarily of a technical nature; the BACnet, KNX, MQTT or OPC-UA protocols exist. The challenge is structural: who is allowed to share which data with whom, under what conditions and for what purpose? This is the question of data sovereignty, which is still largely unresolved in existing buildings.

Sources: Fiyam Digital (2025): ROI analysis predictive maintenance buildings; Baumeister (2025): “Predictive Maintenance: Intelligent Buildings”; Vodafone Immobilienwirtschaft: “Predictive Maintenance for the Real Estate Industry”; dena (2024): Energy-efficient AI study

What are the biggest hurdles today when it comes to sharing data from building automation, technical systems, ownership structures and user contexts?

The hurdles are multi-layered, and you have to separate them cleanly to avoid falling into the trap of solving a technical problem while the actual blockage is a regulatory or organizational one.

Firstly: technical heterogeneity. The building stock in Germany has grown historically. We have proprietary BMS systems of different generations that often do not offer standardized export interfaces. Although open standards such as IFC or COBie for BIM-based FM are available, they are only implemented in a fraction of the existing buildings. What’s more, the data quality required for AI – validated, complete, time-stamped – is often not achieved without considerable effort, even in more modern installations.

Secondly: ownership and stakeholder structures. In the housing stock, we typically have at least four parties with different interests: Owner, housing association, FM service provider and tenant. Everyone is a data owner in their own area, no one has a complete picture. The question “Who owns a tenant’s consumption data?” has no clear technical or legal answer today and thus blocks AI applications that would need precisely this data.

Thirdly: GDPR and purpose limitation. Even where anonymization would be technically possible, there is a lack of cross-sector solutions. The example of heating systems makes this clear: although autonomous control is already technically possible today, available consumption data is often not allowed to be processed automatically – resulting in unnecessary CO₂ emissions that could be avoided if the data were used in a legally secure manner. This is where we need dedicated, consent-based data release mechanisms as a technical infrastructure, and this is a core contribution of the SmartLivingNEXT approach.

Sources: Facility Management Trade Journal (2023): “Making digital building data more accessible”; BIM Germany: “Predictive maintenance and BIM”; SmartLivingNEXT Quarterly Report Q4/2025

SmartLivingNEXT is working on linking data from different data owners in a controlled, consent-based and GDPR-compliant manner. Why is this a strategic advance for smart buildings?

Because it is the only architecture that makes trust scalable. Sounds abstract, but it’s not. The previous logic was: if you want to share data, you either have to trust a central platform, with all the lock-in risks, or do without data sharing. Both are dysfunctional for the German and European market.

The SmartLivingNEXT approach of the federated-decentralized data room is based on a fundamentally different principle: data remains in its original systems. No central data warehouse is set up. Instead, a standardized semantic data model, the SENSE WoT standard, enables data from different owners to be made usable across systems without giving up their sovereignty over the data. Data providers and data users retain complete control and can explicitly decide whether and for what purpose their data may be used.

This is strategically important because it opens the way for multi-actor applications that were previously not legally or technically feasible. If I, as a housing association, can link my energy data with the metering point operator and the FM service provider for a specific purpose, in compliance with the GDPR, auditable and with clear access rights, then applications will emerge that were simply not possible before. The data room becomes a regulated infrastructure, similar to an electricity grid: open, neutral, but with clear access and usage rules.

The fact that this has also been recognized politically as a key issue is shown by the so-called ‘Berlin Declaration’ of October 2024, in which representatives from industry, the housing sector, science and trade jointly proclaimed the development of secure data rooms for the complete digitalization of residential buildings as a national goal.

Sources: SmartLivingNEXT (2024): Berlin Declaration, Days of Digital Technologies; SmartLivingNEXT kick-off event: Thomas Feld, Materna; EU Data Act (Regulation (EU) 2023/2854), in force since 11.01.2024, applicable from 12.09.2025

What will be possible for the first time in existing buildings if data from the building envelope, from technical systems and, if necessary, from the private usage context of an occupant can be merged in a legally compliant manner?

This is the moment when a ‘data silo ensemble’ becomes a learning system. In concrete terms, four application classes open up in the inventory that are either not feasible today or can only be realized in fragments:

Firstly, genuine energy optimization in terms of building physics and technical systems: if heat transfer coefficients from the energy performance certificate, current heating curve data and individual usage profiles of a household – for example, absence times and ventilation behaviour – are fed together into an AI model, heat demand forecasts can be made at building level that go far beyond today’s heating automation. The energy efficiency data portal (EEDP) in the SmartLivingNEXT context shows that reliable heat forecasts can be derived on this basis in the early project phases.

Secondly, proactive, condition-based maintenance: instead of reactive maintenance according to a schedule or after a breakdown, condition data from building operation can be linked with manufacturer data on system components and external climate data. Elevators, heating circuits or ventilation systems become monitored systems whose remaining service life is continuously recalculated.

Thirdly, dynamic energy flexibility management: the so-called redispatch concept, which is being developed in the SmartLivingNEXT data room, uses cross-building consumption forecasts to detect grid overloads at an early stage and avoid peak loads. This turns buildings into active participants in the energy system – a paradigm shift.

Fourthly, personalized, comfort-enhancing services: When residents voluntarily contribute their private usage context – and receive measurable added value in return – assistive services are created that support self-determined living. This is particularly relevant for the growing group of older people, for whom living in existing properties is often the alternative to inpatient care.

Sources: SmartLivingNEXT Quarterly Report Q4/2025; SmartLivingNEXT (2023): Kick-off event, energy use case; BIM Germany: Predictive maintenance; German Research Center for Artificial Intelligence (DFKI)/SmartLivingNEXT: AI basic services for energy data

Many debates about AI in the building sector remain abstract. Which applications will become more concrete and economically attractive through such governance-enabled data linking?

The debate remains abstract precisely when AI is treated as a technology issue and not as a business model issue. The crucial question is always: what economically quantifiable added value does an AI application generate, for whom and at what cost?

Let me name four applications that gain significantly in efficiency through governance-enabled data rooms: Firstly, automated operational optimization. Since the end of 2024, the GEG has stipulated minimum requirements for building automation for almost all non-residential buildings in Germany. AI-based solutions for predictive control, i.e. demand-based control based on weather forecasts and room usage data, will therefore no longer just be desirable, but a regulatory requirement. Governance-capable data connections make these solutions portable between different buildings and operators.

Secondly, ESG reporting automation. The EU taxonomy, the CSRD and the EPBD generate enormous data requirements in portfolio operations. Anyone still manually compiling consumption data today will not be able to scale ESG reporting. Governance-enabled data rooms enable the automated aggregation of valid building data for reporting, a direct business case for portfolio owners and asset managers.

Thirdly, demand side management and dynamic electricity tariffs. The FAME4ME project in the SmartLivingNEXT network is investigating how AI-supported platforms can communicate individual dynamic electricity tariffs and thus address user groups in a targeted manner. Without reliable usage data – which is only possible with consent-based data access – these tariff models cannot be implemented fairly and effectively.

Fourthly, smart insurance: early damage detection across buildings using IoT and AI is a growing market for InsurTech companies. These services can only be rolled out to the portfolio with a standardized, governance-compliant data situation.

Sources: SmartLivingNEXT/Fraunhofer ISE: FAME4ME project; aedifion (2024): “Trends in building management 2024”; SmartLivingNEXT market study 2025 (German Federal Ministry for Economic Affairs and Climate Action (BMWK); GEG requirements for building automation in non-residential buildings 2024

Is this governance capability ultimately the actual prerequisite for AI to become resilient and scalable in existing buildings?

Yes, and I would go even further: Without robust data governance, AI in existing buildings is not structurally scalable. This is not a technological statement, but a logical one.

AI models for building operation, whether for energy forecasts, failure predictions or user behavior, need three things at the same time: sufficiently large training data sets, sufficiently diverse data sets and continuously updated operating data. None of these three criteria can be met in a single building. A single apartment building with 30 units simply does not generate enough data points to train a reliable ML model. Only when I can aggregate data from hundreds or thousands of buildings, while preserving the data sovereignty of each individual owner, do the training data sets emerge that make AI models truly powerful.

SmartLivingNEXT has addressed this problem with an elegant approach: instead of aggregating real personal data, language models are used to generate synthetic energy data that can be used for AI training without recourse to personal data. German Research Center for Artificial Intelligence (DFKI) has shown that this approach is very well suited to realistically simulating various household types, resident structures and appliance combinations. This is governance design on a technical level, and it is the only way to take the GDPR and the data needs of AI seriously at the same time.

Governance capability does not mean regulatory bureaucracy, but rather the opposite: it is the infrastructure that enables trust to arise, and thus data sharing and scalable AI. The predecessor project ForeSight has already shown that AI is possible in residential buildings if a sufficiently large pool of data is available. This is precisely where SmartLivingNEXT comes in and creates the infrastructure for this.

Sources: SmartLivingNEXT Quarterly Report Q4/2025 (German Research Center for Artificial Intelligence (DFKI), synthetic data); SmartLivingNEXT: Days of Digital Technologies 2024, ForeSight award; EU Data Governance Act (DGA), valid since 24.09.2023

What is the significance of such an approach for European and medium-sized providers who want to develop innovative services without being dependent on closed platforms?

In my view, this point is one of the most important strategic aspects of the entire project and it is still not sufficiently recognized in the public debate.

The alternative to an open, federated, neutral data ecosystem are closed platforms from large technology groups. These already exist: hyperscalers from the USA and proprietary building automation manufacturers are trying to gain data sovereignty in buildings via their cloud platforms. For small and medium-sized providers, start-ups, regional energy service providers and specialized FM software houses, this means either dependency or exclusion from the market.

SmartLivingNEXT’s ‘best of both worlds’ strategy is a structurally important counter-model here: manufacturers and providers do not have to fundamentally change their technology or business models. The SmartLivingNEXT ecosystem adds new and AI-supported services that can scale quickly and securely across a large building stock. Access to the data room is standardized, manufacturer-neutral and undisclosed for existing system environments.

This should also be seen in the light of the EU Data Act (in force since January 2024, applicable since September 2025), which explicitly addresses lock-in effects in cloud services and prescribes the interoperability of cloud services. Combined with the Data Governance Act, which regulates data brokerage services and aims to strengthen trust in voluntary data exchange, a European legal framework is being created for the first time that favors open, federated ecosystems over closed platforms. Those who invest now in GDPR-compliant, DataAct-compatible infrastructures are positioning themselves for the next decade – and this is particularly true for European SMEs, which have real technological expertise, especially in niche applications.

Sources: EU Data Act (Regulation (EU) 2023/2854, applicable 12.09.2025); Data Governance Act (DGA), valid 24.09.2023; SmartLivingNEXT Summit 2024: Michael Schidlack, FE at ZVEI; dawex.com (2024): “In-depth insight into DGA and Data Act”

What is the most important message for you with regard to the coming years in the smart living and building sector?

My most important message is that the central bottleneck no longer lies in the technology. It lies in the structural connection between existing systems and for this we need governance infrastructure, not new devices.

We are at a turning point. SmartLivingNEXT will go live when it is completed in August 2026. This is not the end of the project – it is proof that a manufacturer-neutral, AI-enabled data ecosystem for residential buildings works from a technical and regulatory perspective. The Berlin Declaration of October 2024 is the political signal that industry, the housing sector, science and trade want to go down this path together.

What is needed now are three things: firstly, scaling – the developed technologies and governance structures must be transferred to real market operation. Secondly, standardization – the SENSE WoT standard and open semantic models must become the industry norm, similar to IFC/BIM in the planning phase. Thirdly, business model development – the community of satellite projects and associated partners must develop viable digital services in the ecosystem that convince tenants, owners and operators alike.

The heating transition, the energy transition, ESG obligations and demographic change – none of these issues can be resolved in existing buildings without digitalization. But digitalization alone is not enough. What we need is digitalization with data sovereignty. And that is the precise interface that SmartLivingNEXT is working on. The next phase is the market and I am convinced that European providers can play a leading role in this if we make the right infrastructure decisions now.

Sources: SmartLivingNEXT closing event (planned for 2026): “From research to real operation”; SmartLivingNEXT market study 2025 (German Federal Ministry for Economic Affairs and Climate Action (BMWK); Berlin Declaration (08.10.2024), Days of Digital Technologies

About the interviewee

Prof. Dr. Christian Schlicht is a professor and recognized expert for data and artificial intelligence in the construction and real estate industry as well as facility management. His research and consulting work focuses on the digital transformation of the building stock, data-based operational optimization and the governance requirements for scalable AI use in the real estate sector.