The Customer
Associa
The Project
AI Project
Overview
Associa, the largest HOA management company, tasked Improving with enhancing their widely used Town Square app. The primary focus was to improve the resident request feature, which was frequently complained about. The goal was to leverage AI to assist managers in handling these requests more effectively, thereby improving the user experience for both managers and residents. The project involved developing a human-in-the-loop AI solution to streamline the response process.
The Business Challenge
Associa faced significant challenges with the resident request feature in their Town Square app. This feature, though heavily used, was also the most complained about. Managers were struggling to handle requests efficiently, leading to slow and sometimes non-existent responses. This inefficiency not only affected user satisfaction but also hindered the overall effectiveness of property management operations.
Our Solution
Improving approached the problem by developing a human-in-the-loop generative AI solution. This solution aimed to assist managers in crafting timely and accurate replies to resident requests. By using a combination of AWS Bedrock, tool routing agent architecture, and retrieval-augmented generation, the AI system could generate responses based on community documents. Managers were given the flexibility to use, modify, or discard these AI-generated replies before sending them to residents.
The Business Benefits
Improved efficiency: Managers can now handle requests faster, reducing response times from 45 minutes to just 5 minutes.
Enhanced user experience: Residents receive quicker and more accurate responses, improving their overall satisfaction.
Cost savings: The AI system reduces the time managers spend on mundane tasks, allowing them to focus on more critical responsibilities.
Scalability: The solution is designed to scale, with plans to increase traffic from 1.1% to 100%, accommodating all communities managed by Associa.
Market differentiation: By integrating AI, Associa can leapfrog competitors and offer advanced, cost-effective services.
Data-driven insights: The use of tools like Jupyter notebooks and Metabase enhances data-driven decision-making processes.
Technologies and Methodologies Used
AWS Bedrock: Ensured all data remained within AWS, facilitating smoother integration and compliance.
Anthropic Cloud: Provided the AI models necessary for generative tasks.
Langchain: Used for tool use routing and retrieval-augmented generation.
MongoDB: Leveraged for existing data storage within the Town Square app.
Jupyter Notebooks: Enabled detailed analysis and presentation of data.
Metabase and Redshift: Utilized for data-driven insights and reporting.
Partnerships
Improving worked closely with AWS, which provided critical support and infrastructure for the project. AWS's involvement was instrumental in overcoming scaling challenges and ensuring seamless integration of AI technologies. The collaboration with AWS ensured that the project stayed within compliance parameters and received timely assistance when needed.
Lessons Learned
Design pilots meticulously: Ensuring that pilot phases are well-planned and data is thoroughly analyzed is crucial for success.
Understand cost dynamics: The cost of LLM tokens is minimal compared to infrastructure costs, highlighting the need to focus on storage and compute expenses.
Adopt a data-driven approach: Leveraging data for decision-making is essential in AI projects to achieve accurate and efficient results.
Human-in-the-loop reduces risk: This approach mitigates risks associated with AI implementation by keeping human oversight in the loop.
Effective partnerships: Collaborating with partners like AWS can provide vital support and resources, ensuring project success.
Scalability considerations: Anticipating and planning for scaling challenges is necessary to ensure smooth transitions from pilot phases to full-scale implementation.
Conclusion
The AI Project with Associa showcases Improving's ability to leverage cutting-edge technology to solve complex business problems. By developing a human-in-the-loop AI solution, we significantly improved the efficiency and user experience of the Town Square app. Our strategic approach, combined with strong partnerships and a data-driven methodology, provided Associa with a scalable and competitive solution. This project highlights our expertise in delivering innovative AI solutions that drive tangible benefits for our clients.
