The Housing Choice Voucher (HCV) program, funded by the Department of Housing and Urban Development (HUD) and administered by state and local public housing authorities (PHAs), provides essential rental assistance to low-income families, the elderly, and the disabled.
The program is important for more than 2 million households, but a confluence of pressures challenges its administration. These barriers include a complex framework of federal regulations, local PHA policies and procedures, and the need to manage multiple systems for data, document management, workflow tracking, appointment scheduling and reporting.
Additionally, reduced funding, staff turnover and stringent time constraints further complicate the environment, often resulting in backlogs and delays in responding to inquiries from key stakeholders such as applicants, participants and landlords.
Implementing an AI-powered virtual assistant
Artificial Intelligence (AI) virtual assistants offer a promising solution to enhance program efficiency by supporting staff, participants, applicants, and landlords in navigating PHA policies and HUD regulations.
To help PHAs overcome these challenges, CGI developed the Electronic Virtual Assistant (EVA), a retrieval augmented generation (RAG) chatbot. EVA integrates HUD's Housing Choice Voucher Guidebook, PHA-specific administrative plans, federal regulations, and procedure manuals to provide precise answers to complex inquiries, enhancing decision-making, operational efficiency, and program compliance. This eliminates the need for staff to spend hours or days manually researching answers, reduces the risk of providing incorrect information, and thereby improves accuracy.
EVA empowers PHA staff to quickly address questions during calls, or while processing other HCV program tasks such as recertifications, eligibility determinations, lease-ups and housing assistance payments. Seamlessly integrating with existing systems, EVA enhances process workflow without disrupting current operations.
Additionally, PHAs can deploy EVA as a public-facing agent to handle routine inquiries from participants, applicants, and landlords, while escalating complex issues to human staff. This approach reduces call center burdens and provides the public with quick, accurate answers, improving customer satisfaction and operational efficiency.
By leveraging EVA, PHAs can expect to see significant reductions in average call handling times and improvements in program compliance. The AI assistant pays for itself in Year One and delivers up to six times its value in return on investment over 5 years, while improving compliance and participant experience.
An AI-powered virtual assistant, equipped with PHA policy and HUD regulations, transforms the HCV program by boosting communication, accessibility and operational efficiency. By strategically overcoming challenges, this technology empowers staff, participants, applicants and landlords to quickly find accurate information, ensuring the mission of providing affordable housing is met.
Benefits of EVA for the HCV program
- Policy and regulation expertise: AI virtual assistants can interpret PHA policies and HUD regulations, providing accurate information to stakeholders and reducing staff workload.
- Streamlined communication: AI can answer common questions and guide staff, participants, applicants and landlords on program requirements.
- 24/7 accessibility: AI offers round-the-clock support, enabling stakeholders to access information and assistance anytime, enhancing accessibility.
- Consistency in information: AI delivers consistent responses, reducing misinformation and building trust among stakeholders.
- Efficiency in operations: AI automates routine inquiries, allowing staff to focus on complex cases and strategic initiatives, boosting program efficiency.
Avoiding pitfalls
- Challenge: Limited human Interaction → Solution: Implement a hybrid model where AI handles routine inquiries and human agents address complex or sensitive issues, ensuring personalized support when needed.
- Challenge: Privacy and security concerns → Solution: Develop and enforce strict data protection protocols, including encryption and regular security audits, to safeguard sensitive information. Educate stakeholders on data privacy practices.
- Challenge: Complicated implementation → Solution: Conduct a phased rollout of AI technology, starting with pilot programs to test functionality and gather feedback. Provide comprehensive training for staff to ensure smooth integration.
- Challenge: Potential bias → Solution: Regularly audit AI systems for bias and fairness, using diverse datasets for training. Involve stakeholders from various backgrounds in the development process to ensure inclusivity.
- Challenge: Dependence on technology → Solution: Establish contingency plans for technical failures, including backup systems and manual processes. Ensure continuous monitoring and maintenance of AI systems to minimize disruptions.
