with expert Stefan Cadosch, former SIA president and CEO of AI-startup keeValue AG
Summary
In this episode, Johannes Lohner talks with Stefan Cadosch – architect, former president of the Swiss Society of Engineers and Architects (SIA), and now CEO of Key Value AG. The discussion focuses on how building data and artificial intelligence (AI) can help provide highly accurate construction cost forecasts, increasing planning and investment security. Other key topics include life cycle analysis, the use of digital tools in practice, and real-world experience from Switzerland. The podcast offers valuable insights for developers, planners, and decision-makers in the construction industry.
The Power of Building Data in Early Project Phases
Stefan Cadosch explains the great uncertainty that developers face in early planning phases—often based on rough estimates. Key Value uses a large dataset of completed projects to support architects and planners with accurate figures very early on. Instead of relying on wide cost estimation ranges, data-based tools provide realistic values for construction costs—often with less than 10% deviation. Especially for new builds, this allows a high level of reliability already at the “cost estimate” stage. The foundation: as much high-quality, real-world building data as possible.
AI-Based Forecasts and Cost Certainty
A central part of the conversation is the use of artificial intelligence (AI) for predicting construction costs. Key Value’s algorithms are trained using actual cost reports—not theoretical models. Typical surprises in renovation projects, for example, are already reflected in the data. This results in remarkable accuracy: renovation projects currently show a forecast precision of around 12–15%, far better than conventional methods. One multifamily housing project calculated with the tool ended up with only a 0.3% deviation. Stefan emphasizes: “The curse of the first number”—that first rough cost estimate—sticks with the client forever. This is where AI provides reliability from the start.
Planning Certainty Through Digital Tools
Using typical project phases, Stefan explains how cost estimate accuracy evolves over time—and where conventional approaches fall short. Key Value introduces the precision of a detailed cost forecast at the early “preliminary project” stage, helping avoid flawed planning or costly project cancellations. Especially valuable is the integration with BIM models: with every design change—like adding a wall—architects immediately see the impact on costs and CO₂ footprint. This enables a holistic decision-making framework that balances design, ecology, and economics—a real shift in planning culture.
Life Cycle Analysis and Investment Strategy
Another major topic is life cycle analysis. Stefan explains that most building costs occur not during construction, but throughout decades of operation. Key Value has developed tools that calculate all costs—from the first sketch to demolition—by year and building element. This allows strategic planning of renovations and smart bundling of investment decisions. A new feature is the life cycle assessment tool, which links environmental impact (e.g., CO₂ emissions) directly to cost. With this, users can compare variants and find the right balance between sustainability, budget, and design—“Design to Cost to Ecology.”
From Switzerland to the World – and the Limits of Data-Driven Systems
In the final part of the episode, the discussion turns to the application and scaling of the tools. Key Value’s database is especially strong for residential and public buildings. For niche types like hospitals or rarely built structures, predictions are more difficult. Still, regional differentiation is impressive: the same project yields different cost results in Zurich, Bern, or Ticino—reflecting local market conditions. The team is now expanding into Germany, facing challenges like different standards, terminology, and building cultures. But with each new country, the system gets smarter. The goal is a platform that scales internationally—always based on real-world building data.