Contact
Scout24

Meet HeyImmo: Your Personal AI Assistant for Real Estate Questions.

Immo Scout24
TagsCustomer ServiceRetail & E-CommerceMedia & EntertainmentAgentic AI
Immo Scout24
TagsCustomer ServiceRetail & E-CommerceMedia & EntertainmentAgentic AI

Scout24 is one of Germany’s leading technology companies, connecting millions of people with residential and commercial real estate through ImmoScout24. Together with ML6, Scout24 further developed HeyImmo, the AI-powered real estate assistant for the ImmoScout24 platform. HeyImmo helps users with concrete questions around real estate, including property search, valuation, financing, renting, buying, and individual listings. The assistant is designed to make complex real estate decisions easier to navigate through clear, goal-oriented interactions.

Find out more
Scout24_case study

Catch up’ quickly

Scout24 enhances real estate journeys with HeyImmo, an AI-driven assistant that improves user satisfaction, response speed, and support quality for millions navigating property transactions.

Continue reading

About this client

Scout24 is one of Germany’s leading digital companies and operates ImmoScout24, the market-leading platform for residential and commercial real estate. For more than 25 years, the company has connected homeowners, tenants, buyers, and real estate professionals, helping millions of people navigate one of life’s most important decisions: finding, renting, buying, or selling property. Today, around 19 million users visit ImmoScout24 each month. Beyond property search and listings, Scout24 is continuously expanding its digital ecosystem with products and services that support the entire real estate journey. Through data, technology, and innovation, the company is helping to make real estate transactions more transparent, efficient, and digital across Germany, Austria, and Spain. Scout24 is a listed stock corporation (ISIN: DE000A12DM80, Ticker: G24) and member of the DAX, the DAX 50 ESG and the DAX 50 ESG+.

impact

By improving response quality and reducing latency, the solution delivered a noticeably better user experience. Users found answers faster, trusted the responses more, and reported higher overall satisfaction.

75%

User Satisfaction, up from 68%.

85%

Quality Score, up from 80%, based on internal quality evaluations.

5s

Time to First Relevant Token (TTFRT). Down from 8 seconds, meaning users see a relevant response 3 seconds sooner.

visual 2 - scout24

Challenge

Finding, financing, renting, buying, or selling a home is rarely a linear process. Users often need guidance across several steps, from understanding a listing to comparing options, preparing a valuation, or finding the right next action. Scout24 introduced HeyImmo to make this journey easier and more intuitive. As the assistant moved from innovation project to scaled product experience, several challenges became central.

  • Guiding complex user journeys

    Users needed more than static search results. HeyImmo had to understand the user's goal and guide the conversation in a helpful, focused way.

  • Response latency

    To feel natural and keeping users engaged, the assistant had to respond quickly and reduce waiting time during the interaction.

  • Managing relevant context

    HeyImmo draws from four distinct data sources simultaneously. HeyImmo needed to use the right information at the right moment without adding unnecessary complexity or latency.

  • Scaling personalized support

    The experience had to support a growing user base while maintaining answer quality, reliability, and performance.

  • Integrating across a broad product landscape

    The solution needed to work across different product areas and data sources while remaining maintainable for the teams involved.

  • Maintaining quality at scale

    As the assistant became more capable, Scout24 needed a structured quality framework to help ensure helpful, accurate, and safe responses.

SOLUTION

The Technology
behind HeyImmo

HeyImmo is designed to make complex real estate questions feel easier to handle. The user experience is intentionally simple, while the underlying setup supports reliability, speed, and quality at scale.

  • Powered by State-of-the-Art LLMs

    HeyImmo uses advanced language-model capabilities to understand user intent and keep context across multi-step real estate journeys.

  • Goal-oriented guidance

    HeyImmo does more than answer isolated questions. We’ve moved away from rigid "if-this-then-that" logic to a dynamic goal-steering system. This means the AI helps users move toward a concrete real estate goal, such as understanding a listing, preparing for a valuation, or finding the next relevant step.

  • Built for speed

    Nobody likes waiting for a screen to load. We’ve re-engineered the architecture to remove "digital clutter," ensuring responses are faster and make interactions feel more fluid.

  • Smart Thinking, Less Waste

    To keep the experience lightning-fast, HeyImmo only "loads" the context information it needs exactly when it needs it. This on-demand intelligence reduces lag and ensures the system remains lean and efficient and users receive relevant answers without unnecessary system complexity.

  • Professional Quality Assurance

    The project integrated a rigorous evaluation framework that combined human manual testing with automated LLM-as-a-Judge (LaaJ) metrics. With a quality target of 85% or higher, we ensure every interaction feels as reliable as speaking with a seasoned real estate professional.

Strategy,
Governance
and Scale

HeyImmo is more than a chatbot. It is part of Scout24's broader approach to applying AI in ways that are useful, maintainable, and scalable.

The platform uses the Model Context Protocol (MCP) to allow multiple product teams at Scout24 to integrate their own specialized tools and data independently. Standardized MCP guidelines and upskilling sessions ensure all teams build consistent, high-quality integrations.

Tech Stack & Services

  • Cloud Infrastructure

    Hosted on Scout24's AWS environment.

  • Model Providers

    OpenAI direct API

  • Backend Frameworks

    Pydantic agent framework

  • Monitoring & Observability

    DataDog

Working with ML6 helped us advance HeyImmo both technically and methodologically. Their deep expertise in building production-grade AI systems, combined with their full commitment to this project, was instrumental in creating an assistant that delivers value to our users.
Daria Skachek
Daria SkachekEngineering Manager of HeyImmo

Results

  • Optimized Output Quality and Reliability

    The implementation of advanced conversation steering, centralized prompt management, and a structured QA process directly enhanced the AI's performance. Internal quality evaluations rose from 80% to 85%, ensuring that guidance provided to users regarding property search and valuation is consistently accurate and reliable.

  • Faster, More Natural Interactions

    By re-engineering the underlying architecture and removing redundant processing steps, response latency dropped significantly. Time-to-first-token improved from 8 seconds to under 5 seconds, transforming the experience from a traditional search query into a fluid, human-like conversation.

  • Enhanced User Experience and Satisfaction

    Refined interaction logic and faster responses helped increase user satisfaction from 68% to 75%.

  • Scaled Market Adoption and Engagement

    Strategic technical optimizations enabled HeyImmo to support a growing volume of user interactions while maintaining performance and quality. The assistant has established itself as an integral part of the ImmoScout24 experience, demonstrating its ability to deliver value at scale.

Fabian Schrempf 7 (1)
HeyImmo is a prime example of what becomes possible when deep AI expertise meets an ambitious partner. Together with Scout24, we did not just build an assistant, we built a scalable AI platform that guides millions of people through one of the most significant decisions of their lives.
Dr. Fabian SchrempfHead of GTM DACH

The Collaboration

The development of HeyImmo followed a hybrid co-creation model. Scout24 contributed deep real estate expertise, product ownership, and platform knowledge. ML6 supported the project with applied AI and machine-learning expertise.

The collaboration was designed not only to deliver a product improvement, but also to strengthen Scout24's ability to operate, evaluate, and evolve AI-powered product experiences independently.



Stocksy_txp813879ba5En200_Small_676624
Share this on:
Sebastiaan Tuts 12
Sebastiaan TutsHead of Business Development

Inspired?
Let’s connect and make it happen!

Ready to elevate your AI game? Schedule a meeting with us today and let’s craft a winning strategy together!

Let's schedule a talk
Sebastiaan Tuts 12
Sebastiaan TutsHead of Business Development