Pioneering work: Artificial intelligence is also ushering in a new era in real estate management
Artificial Intelligence (AI) is a groundbreaking technology that has the potential to revolutionize all aspects of life. While it offers unprecedented opportunities, it also raises fears, such as being perceived as a potential threat to job security. Deutsche Telekom faces a unique challenge in this context: Can AI contribute to saving 10% of operational costs annually? How can it be ensured that this technology is seen as valuable support to employees rather than a threat?
We gladly took on this challenge, which aligns perfectly with our defined purpose: to employ new technology to benefit people and society.
How to Increase Cost Efficiency if You Have Almost Infinite Possibilities?
Deutsche Telekom owns several thousand buildings in Germany alone, resulting in an annual operational cost of over one billion euros. Therefore, real estate management challenges include optimizing this vast property portfolio, where even small efficiency gains can lead to significant savings. The general aim is to maximize building occupancy and reduce inefficient space use.
There are many factors to optimize around, making this a complex task. Buildings can be owned or leased, with various lease durations and contractual aspects to consider. The building's location, infrastructure, user structure, and specific unit requirements – must be considered while keeping the available space in mind. For mere humans, the number of possibilities is overwhelming, amounting to a number with 602 digits, a scale utterly incomprehensible to us. However, this is an ideal scenario for applying "Machine Learning."
The AI Recommends, The Expert Decides
From the outset, it was clear that AI should function not as a replacement but as a support to human expertise. Machine learning can better handle rational aspects and evaluate more options in less time. However, AI cannot replicate the experiential knowledge from direct interaction with building users and the emotional intelligence of experts. Employees should be empowered to assess AI-proposed scenarios to increase building efficiency and occupancy and make informed decisions to save 10% of operational costs.
Involving Users in "Co-Creation" in Development
The implementation began with carefully analyzing existing workflows in Real Estate Management. Through workshops and interviews with employees, we understood how current processes work, their challenges, and potential needs that could benefit from the new solution. These insights were crucial to developing a tool that seamlessly integrates into daily workflows, combining human intuition with machine efficiency.
The design of the User Interface was of central importance. It needed to be intuitive, visually appealing, and tailored to user needs. Hence, intermediate steps were regularly validated with future users and continuously developed. We supported its introduction with a comprehensive training program to ensure employees felt comfortable with the new tool and could fully exploit its potential.
The final tool was designed around typical user workflows, incorporating unique features to enhance user satisfaction, like export options for graphical relocation plans.
The tool works as follows: Real Estate Managers first define a scope, specifying the number of buildings to be optimized. Employees set essential parameters, based on which the AI determines all possibilities and evaluates them in different scenarios. The system then creates a prioritized list of these machine-evaluated scenarios. Employees select those that seem realistic to assess their feasibility in more detail and consult with tenants, other parties, departments, or stakeholders.
Technical Implementation and User Acceptance
The technical implementation involved developing a powerful AI system (Backend) capable of conducting complex data analyses and suggesting efficient usage scenarios for several properties. This development was undertaken by an internal team of experts at Deutsche Telekom. The system was designed to consider various parameters such as location, user structure, and lease conditions and generate optimization suggestions based on these. The User Interface (front-end) was developed with our development partner, Nerdgeschoss, using a design system we created.
The new tool has been very positively received
Users rate it with an NPS of 57 and assess its usability (SUS) at 72%. They also rate the fulfillment of desired tasks up to 85%. The AttrakDiff value, a measure of product attractiveness, classifies the tool as "Desired," the highest rating, confirming its successful integration into daily work and high user acceptance.
This project impressively demonstrates how close collaboration with end-users and integrating their needs and ideas into the development process results in a highly functional and user-friendly tool that achieves immediate acceptance and application.
User Acceptance is Key to Product Success
User acceptance of a new solution is crucial for product success. This is especially true in this project, where the user group is relatively small and cannot be expanded through marketing activities. Involving users in the development process is essential when developing such specific solutions. Our Co-Creation methods have significantly affected acceptance and readiness for use.
Artificial Intelligence for the Benefit of People and Society
This project demonstrates that when AI is carefully implemented with consideration for people's needs and concerns, it can be a powerful tool for efficiency enhancement and cost reduction. It also shows that integrating AI into work processes does not necessarily lead to job loss but can enrich and complement human work.
We Can Help You Involve Users in Development
AI is an excellent opportunity to enrich the work world while keeping human needs at the forefront. This project exemplifies how technology can benefit society by supporting and complementing human work rather than replacing it. However, involving users in the development process is crucial to success. This applies to almost every AI-related project. We are happy to support you with our methodological expertise in "Co-Creation."
