Building property management–specific AI using AWS with Aareon

Aareon seeks to provide complete solutions, empowering its customers with an environment that helps them accomplish all property management processes in a single place. Built for managing large numbers of properties simultaneously, Aareon’s system handles workflows for apartments, payments, maintenance, invoices, energy consumption, and more. It also includes a portal that digitizes work orders, invoices, and scheduling for contractors who do maintenance and repairs.

In the rise of AI, Aareon saw the opportunity to automate some of these workflows and reduce manual tasks for property management companies. “As a human person who uses AI, you can accelerate your daily business routines so that you have more time for the more difficult tasks and decision-making,” says Martin Habib, head of development housing at Aareon in Germany. “Using AI will soon become the normal way of working.” To integrate industry-specific AI capabilities into its environment, the company decided to draw on the high availability, security, and flexibility that AWS offers.

Using AWS, Aareon integrated its virtual assistant, AAVA, in its property management system. The company connected the solution’s backend to Amazon Bedrock, a service for building generative AI applications and agents at production scale. This way, Aareon can use large language models (LLMs) while helping to meet General Data Protection Regulation (GDPR) security and privacy requirements. Any international data transfers are protected by data processing agreements and appropriate safeguards under the GDPR. In addition, customer data is not used to train or improve AI models.

This helps the company secure data and comply with the GDPR under the AWS Shared Responsibility Model.

Aareon uses various LLMs in Amazon Bedrock, including Anthropic’s Claude models and Amazon Nova models—fast, cost-effective models for production AI applications. The company can test LLMs and select the model with the optimum accuracy and price performance for a specific use case. For example, Aareon uses a model with a higher token limit for calculating service charges, which is a more complex use case. “By connecting to Amazon Bedrock, we have access to a lot of LLMs that are ready to use,” says Martin Habib. “And when new versions arrive, we can test them out. We can decide on a model that matches the complexity of the use case through trial and error. That flexibility is valuable.”

Besides flexibility, Aareon chose Amazon Bedrock for its high standards of security and its availability in the AWS Germany Frankfurt Region. Using AWS, the company can provide its customers reliable uptime in addition to security and data protection, which are particularly important for companies with large amounts of tenant data. “We wouldn’t choose Amazon Bedrock if we had any doubt about the security and data protection that we could achieve by using it,” says Martin Habib. “AWS specializes in security and could give us advice on improving our posture.”

With support from the AWS team, Aareon has increased its team’s capabilities and implemented concrete solutions. The company also accelerated the project by working with Reply, an AWS Partner, to build the solution interface. Aareon developed the first version of AAVA in 4 weeks and announced it to customers in 6 months. “I’ve never experienced a solution that shipped that quickly,” says Martin Habib.

Integrated into Aareon’s German Wodis Yuneo ERP and further connected solutions, AAVA helps customers query documents, request information through chat, automate processes, or accelerate workflows. The assistant can perform tasks such as scheduling appointments between tenants and craftsmen, coding invoices, analyzing images to draft damage reports, and sorting incoming email.

AAVA can handle both text and voice support. And Aareon continues to trial new use cases, testing which functions work well with voice, text, or image recognition and which models suit each use case. Trained for industry-specific scenarios, the models provide one of the first implementations of AI specifically for real estate management.

Using AI on AWS, Aareon has accelerated its development lifecycle. As the company continues to innovate, it’s looking to help customers analyze data and use AI for fraud and anomaly detection. Aareon will also continue to speed up the automation of time-consuming routine workflows, such as service charge calculation. Overall, the company’s AI solutions help reduce repetitive work. This frees up customers to focus on meaningful customer interactions while building more resilient, future-ready organizations that can better address ongoing workforce shortages.

“AWS provides technology and, more importantly, empowers its customers to use that technology, to have fun with it, and to be successful,” says Martin Habib. “We were supported by and could accelerate the whole process using the expertise that AWS offers. And on the other side, we uplifted our products and product development to deliver valuable AI features to our customers.”