At Mono, we’ve been steadily pushing the boundaries of how AI can genuinely transform business operations. If you’ve read our previous chatbot case study, you know that our earlier chatbot was already great at quickly providing tenants with accurate, property-specific information. But we knew we could push further—moving beyond mere knowledge retrieval to create something much more dynamic and impactful.

Why Agents?

Our original chatbot was effective, but it primarily relied on a RAG (Retrieval-Augmented Generation) approach, meaning it could retrieve and answer questions from client’s internal knowledge base. As client collected enough documents on its rules, processes and procedures, this solution became powerful, but still somewhat closed.

Enter our completely redefined solution: a multi-agent system. By introducing specialized AI agents, we gave our chatbot true actionability. Now, it doesn’t just answer questions—it interacts directly with external services and APIs. It can handle tenant tasks in real-time, whether checking loan or balance statuses, coordinating maintenance requests, or managing parking reservations seamlessly within the client’s existing mobile application.

Tackling Real-World Complexity

Property management involves juggling numerous ongoing demands: rapid response to maintenance issues, effective management of tenant requests, and ensuring non-technical users find their way effortlessly around an app. Traditional solutions were human-intensive, inefficient, and error-prone.

Our new AI-agent-based system solves these problems by delegating specific tasks to dedicated agents. Each agent works independently yet collaboratively, creating a seamless user experience:

• Dispatcher Agent: Acts as the central hub, swiftly understanding and routing tenant requests to the right agent, even handling multilingual communication.
• Maintenance Agent: Expertly captures detailed information on tenant-specific maintenance issues, automates scheduling, and significantly streamlines processes.
• Area Maintenance Agent: Oversees community-wide maintenance tasks, promptly addressing issues across the property.
• Parking Request Manager: Simplifies parking logistics, efficiently managing reservations and vehicle registrations.
• RAG Agent: Maintains accuracy and consistency in providing information from extensive property-specific documentation, complementing other agents.

How We Built It: Framework Selection

Choosing the right foundation was critical. We analyzed several powerful frameworks as potential bases for our agent solution, including CrewAI, LangGraph, and Google’s Agent Development Kit (ADK). Each offered unique strengths:

• CrewAI stood out for simplicity in agent definition and straightforward integration.
• LangGraph provided robust state management and exceptional orchestration capabilities.
• Google ADK offered broad integration options and scalability.

After thorough evaluation, we went with Google ADK as it precisely fits our client’s needs. Other frameworks - or no frameworks at all - are used in our other projects, as there is still no “onw size fits all” solution in this space.

Seamless Orchestration

Underpinning our solution is a sophisticated orchestration layer that manages interactions and maintains seamless handoffs between agents, ensuring a smooth conversational flow. Agents remain contextually aware, leveraging structured data models and multilingual capabilities to ensure accurate, tailored responses.

Real Results

The transition from chatbot to actionable AI agents wasn’t just technical—it transformed operational efficiency and user satisfaction:

• Human intervention in routine tasks reduced by approximately 40%.
• Tenant satisfaction rose significantly thanks to faster, more precise responses.
• Non-technical support inquiries dropped by 50%, significantly improving app usability and reducing support burdens.

What’s Next?

This agent-driven transformation is just a stepping stone. We’re actively enhancing Mono’s AI platform, Dokko, embedding versatile agent functionality that any industry can leverage—not just property management. Future enhancements include integration with telephony systems to enable voice interactions, further extending the system’s accessibility and usability.

At Mono, we’re committed to creating AI solutions that genuinely enhance business operations—moving from static knowledge bases to intelligent agents that act, decide, and truly simplify complexities. Welcome to the next evolution of AI.