Enhancing ADAPT’s Operational Efficiency with Ai: Ensuring Security and Compliance Every Step of the Way
Written by Trang Tran
Background
ADAPT Community Network (ADAPT) is a non-profit organization in New York City, pioneering cutting-edge programs and services for people with disabilities. Their mission is to empower individuals through innovative solutions, one person at a time.
Like many in the healthcare sector, ADAPT faces staffing shortages, making productivity improvement through technology an important task. This challenge has sparked interest in integrating artificial intelligence (AI) into their operations. AI holds immense potential to enhance efficiency but also presents risks such as the possibility of generating false information, along with privacy and security concerns. Furthermore, ADAPT faces unique challenges, including a steep learning curve and infrastructure need for adopting technology, particularly AI, and navigating stringent regulations such as HIPAA and insurance policy requirements. My fellowship represents the first step in addressing these complexities.
ADAPT's Perspective on Applying AI
At the start of the summer, I visited multiple ADAPT sites in Brooklyn, had discussions with ADAPT employees in various capacities, including educational programs for children and adults, residential program, TechWorks assistive technology services, and healthcare clinics. During these visits, I explored the potential of AI applications with staff and participants, discussing practical use cases for the organization.
The potential of AI is promising. It can generate personalized recommendations by analyzing patient data, create tailored content and activities for day and residential programs, and integrate with smart boards or devices to access medical records while tracking individual progress. AI can also automate routine administrative tasks in clinics and offices, such as completing forms, generating reports, and summarizing notes for meetings or doctor visits. In human resources, an AI-powered chatbot could centralize access to information, helping staff quickly find what they need.
Alongside my advisor Ronak Parikh, Senior Vice President of Community & Business Development, we evaluated these AI use cases, weighing both their potential benefits and risks. Based on this assessment, we drafted a policy proposal outlining ADAPT’s approach to AI adoption. Given the rapid evolution of AI, it is important for ADAPT to be proactive while also cautious of associated risks. Our proposed policy covers requirements for AI generated content, emphasizing that all AI output should be proofread, edited, and fact checked by humans. AI-generated content must be transparent, comply with data privacy and fairness guidelines, and include liability coverage for potential risks. Given that ADAPT deals with sensitive information regularly, it is critical that any AI technology applied aligns with data privacy regulations and shields ADAPT and its clients from cybersecurity threats.
AI Education and Training
To support safe AI adoption, I helped design an education and training module for ADAPT employees. This module offers fundamental AI knowledge and specific guidance on the responsible use of AI tools. The training ensures that staff avoid entering sensitive or protected information into unapproved external AI platforms. In addition, for work-related purposes, employees must use only organization-approved AI tools. These are critical measures for safeguarding ADAPT’s cybersecurity.
AI Chatbot for Human Resources (HR)
In response to the challenges faced by ADAPT’s HR department—particularly the difficulty in addressing the high volume of inquiries from staff—I developed an AI-powered chatbot to streamline this process. The chatbot automates responses to frequently asked questions and searches the HR database for information based on user-provided keywords. It responds to questions about employee benefits, company policies, the employee handbook and services, payroll, and the agency schedule. The chatbot can also assist with more general language tasks and redirect users to the appropriate HR staff when more specific support is needed.
Built using Microsoft Copilot, the chatbot integrates seamlessly with ADAPT’s existing infrastructure. The chatbot is adaptable, functioning either as a stand-alone website for employees or as an application within Microsoft Teams, ADAPT’s primary communication platform. By the end of my fellowship, we had tested a demo version that successfully responded to most search queries. However, the chatbot’s answer to knowledge-specific questions still needs improvement, and feedback from the HR department can help to fine-tune the answers further.
Impact and Path Forward
ADAPT's director of IT, Devon Kerr, along with the HR team, is planning to further test and implement the HR chatbot for broader staff use. My advisor Ronak is collaborating with ADAPT’s legal and compliance officers to finalize the AI policy and integrate the AI education module into the organization’s regular training programs. These initiatives will guide ADAPT in responsibly adopting AI while ensuring that security, privacy, and compliance remain priorities. ADAPT is committed to continued exploration of AI opportunities, applying these technologies in safe, efficient, and secure ways to enhance its services and improve the quality of life for the individuals it supports.