Leveling up Building Operations Through Generative AI

Generative artificial intelligence (Gen AI) is a form of artificial intelligence that independently creates new content and solutions based on existing information and spontaneous input — including text, photorealistic images, videos, or audio content. Unlike classical AI, which identifies patterns and solves predefined tasks, Gen AI develops solutions that did not exist before.

Since the breakthrough of ChatGPT, one thing has become clear: technology is fundamentally changing the way we work. The real estate industry is no exception. According to a current trend survey by aedifion and Rueckerconsult, 86 % of the surveyed companies already use AI, and the majority use generative applications such as ChatGPT, Claude, Gemini, or Perplexity.

The real challenge, however, lies in the next step. Individual tools are quick to deploy, but according to the survey, only 27 % of companies systematically integrate AI into their corporate and building processes through platforms that can access their own data sources. This is where the greatest potential lies: combining generative AI with real operational data makes it possible not only to analyze buildings but also to improve them in a targeted way.

In this article, we show which potential generative AI already offers for building operations today and what owners and operators should consider when introducing it.

What Is Generative AI ?

Generative AI goes beyond the applications of classical AI. While conventional systems analyze, categorize, and summarize data, Gen AI generates entirely new content in natural language from existing information and fresh prompts. This includes, for example, text, photorealistic images, videos, or audio content.

This is made possible through Large Language Models (LLMs) that contextualize words and generate coherent text. Multimodal Large Language Models (MLLMs) expand these capabilities by processing and creating multiple data types simultaneously. This makes generative AI a true productivity booster for virtually every economic sector.

According to a study by management consultancy McKinsey, Gen AI technologies could unlock global annual productivity gains worth between EUR 2.4 trillion and EUR 4.1 trillion.

In practice, the trend survey cited above reveals a clear gap: 60 % of participants already use AI for strategic analyses such as portfolio analysis, market research, or ESG reporting. In operational building control, for example in energy management, predictive maintenance, or plant control, the figure is only 23 %. The industry is therefore only partially tapping into the potential available to it. The following examples aim to show where the difference lies between generative AI and "classical" AI, and how owners and operators can make the most of the technology's benefits.

Generative AI vs. Classical AI: What is the Difference?

Classical AI analyzes existing data, identifies known patterns, and solves predefined tasks. In building operations, it can:

• analyze large volumes of operational data in real time,
• perform predictive analyses,
• provide tailored recommendations for action.

Applications range from optimizing materials and building designs to predictive control of heating, ventilation, and air‑conditioning (HVAC) systems and developing new resource concepts. A good example is intelligent demand side management, which uses specific algorithms to manage electricity consumption strategically, reducing costs and stabilizing the power grid.

These conventional applications are valuable but limited in adaptability. So, what if there were a digital building assistant that not only analyzes information but also develops new solutions to ever‑changing operational challenges in real time, directly responding to operator or asset manager queries?

How Generative AI Works in Building Operations

With generative AI, the digital co‑pilot for building operations is no longer a distant vision — it’s already a reality. Here is how it works:

Extensive Data Sources as the Foundation

The foundation is a cloud platform that continuously collects, structures, and processes operational data from the building. The AI assistant is integrated directly into this platform and accesses this curated data base in real time. The model is pre‑trained with massive datasets from public sources, including:

• technical documentation
• operating manuals
• building plans
• sensor data
• databases
• scientific publications
• field reports

This broad spectrum of data gives Gen AI comprehensive domain expertise that is far beyond what conventional AI models offer. The industry increasingly recognizes this potential: 63 % of participants in our survey consider real-time building data important or very important for AI-supported decisions.

"On the technical side, we need live data from the properties. Much of the information from maintenance, inspections, or operations is still collected today with considerable manual effort. AI can make a major contribution, especially when it comes to analyzing this data afterward: we can identify more quickly the condition of a property, how energy consumption is developing, and where action is needed." Thomas Wiese, Technical Property Manager at B&L Property Management

Interactive Communication via Chat or Voice Input

Based on this data, the digital building assistant can respond to so‑called “prompts”. These are natural‑language instructions or questions. Through an intuitive interface such as a chat window, technical operators can ask questions or issue specific commands.

Agent‑based systems autonomously determine which additional data to retrieve for example, the current system status, in order to provide precise, real‑time answers to specific operational questions, not just general ones.

Examples:

• How do I set the correct target value?
• What open tasks remain in the project?
• What are the biggest optimization potentials in the building?
• How can these potentials be realized?

Tailored Recommendations for Optimized Operation

Immediately after receiving a query, the system accesses building‑specific data, activates the relevant information from its training corpus, and returns customized recommendations.

The results are precisely adapted to each building’s individual conditions. If the digital building assistant receives feedback on its responses, it can contextually refine future suggestions.

But facility management isn’t the only field that benefits. Asset managers also gain a fast, user‑friendly way to stay informed about how their portfolios are performing.

Use the Technology Where People Already Work

Modern cloud platforms are no longer closed systems. They act as open data hubs. At aedifion, there are therefore two ways to leverage generative artificial intelligence (AI) in building operations:

1. The aedifion AI Assistant

With the AI Assistant, technical operators and asset managers gain an intelligent co‑pilot that works directly inside the aedifion platform. It is designed for everyone who works there every day, such as technical operators.

Its strength is depth: Users find all fault messages, data points, manuals, project documents, and analyses in one place and receive concrete support for day‑to‑day decisions in real estate operations.

2. The aedifion MCP

MCP stands for Model Context Protocol, an open standard that connects AI assistants to other systems. The aedifion MCP brings data from the aedifion platform into the enterprise AI an organization already uses, such as ChatGPT, Claude, or Langdock. It is designed for everyone who does not always work directly in the platform, such as asset management, ESG managers, or executive management.

Its strength is integration: Responses flow directly into the report, the board memo, or the email to the property manager.

For a general‑purpose AI to interpret building data correctly, the aedifion MCP provides not only the numbers but also the rules for reading them. Example: A green indicator then means “no anomalies detected in the last analysis run,” not “everything is fine.” Control remains with the company. Access is strictly read‑only, and each person sees only the data they are already authorized to access. In short: The aedifion AI Assistant works in our environment. The aedifion MCP works in our customers’ environment.

The examples show how important it is not to view AI applications in isolation, but to bring them to where people already work. This is less about individual tools than about how work processes function as a whole.

"Many people have long used ChatGPT, Perplexity, or other applications to simplify their daily work. The big step, however, is still ahead. We need to make the necessary data available and reshape organizations so that AI makes not only individual employees more efficient, but entire processes. Only then can AI agents, for example, take over structured tasks autonomously." Prof. Dr. Kunibert Lennerts, Professor of Facility Management at the Karlsruhe Institute of Technology (KIT)

Generative AI in Buildings Must Become State‑of‑the‑Art

Whether through the integrated AI Assistant or via the aedifion MCP in the AI tools used by your company, generative AI makes building operations faster, more efficient, and future-proof. Given increasing technical complexity and an ongoing skills shortage, Gen AI is no longer a “nice‑to‑have” — it’s a strategic enabler for smarter decisions, optimized operations, and a competitive portfolio.

The companies in our survey see it similarly: 77 % of AI users plan to use the technology to create better analyses and decision-making bases, and 67 % aim to improve the quality of their services. 50 % name process automation as a goal, and 43 % name innovation and idea development.

"The efficiency gains from AI are so significant that companies must assume employees will want to use such tools. They should therefore evaluate suitable tools as quickly as possible, provide secure solutions, and establish clear rules for data storage and processing." Dr.-Ing. Johannes Fütterer, CEO of aedifion

The key question is therefore no longer whether — but how ready your building portfolio is for AI.