In a previous blog exploring AI, we talked about the usage of AI tools in behavioral health, as well as funding in the state budget for digital therapeutics. While AI has significant application in performing serious tasks such as mental health treatment, there are also basic uses of AI that function on a lower level. This is typically utilized in day-to-day operations. One example is the use of bots in the Ohio Benefits Program, a county-level system that determines program eligibility for services such as SNAP, Medicaid, and TANF. What do these bots do, and how do they assist in service delivery? What do these bots do, and how do they assist in service delivery?
In 2022, the Ohio Department of Administrative Services released a proposal for the usage of AI to assist in workforce development and public assistance programs. The proposal was in partnership with the Department of Job and Family Services (ODJFS), the Department of Medicaid (ODM), and 88 County Departments of Job and Family Services. The collaboration was meant to design, implement, and maintain human service delivery in the Ohio Benefits Program. The bots created in this partnership have multifaceted usability because of the wide range of collaboration in each department.
There are five bots that are currently deployed that have made a tremendous impact on outcomes for service delivery.
Total labor and hours saved: Since implementation, the Ohio Benefits Program Family of Bots reviewed and processed over 500,000 cases, saving caseworkers over five years of working hours. There are five bots that are currently deployed that have made a tremendous impact on outcomes for service delivery.
Deployed Bots in the Ohio Benefits Program
A report released last year, “The Ohio Benefits Program is ‘BOT’ in” provides a description of the various ways that Ohio is utilizing technology tools. This is summarized below.
LTC (Long-Term-Care) Pending Record Removal Bot
The LTC Pending Record Removal Bot leverages process automation technology to clean up data in the Ohio Benefits system. This bot was used to reduce the number of irrelevant long term care records that were created by the customer on the application level. The bot was created in close partnership with county representatives to remove LTC records when services were not needed.
Hours saved: The bot also functions to clean data within Ohio Benefits and has removed approximately 30,000 records from the system.
DRC Bot (Department of Rehabilitation and Correction) Bot
The DRC bot is used to process Medicaid eligibility for incarcerated individuals. This was a project launched in collaboration with ODM and DAS in 2021. Since then, the bot has been used to process new incarceration data from DRC. The bot also updates case information before re-determining Medicaid eligibility and corresponds directly with counties by sending daily reports. The DRC bot receives approximately 4,000 incarceration alerts and can process over 60% of alerts within 24 hours of receiving them. The bot reallocates approximately 2,000 hours of time for workers across the state.
The DRC bot receives approximately 4,000 incarceration alerts and can process over 60% of alerts within 24 hours of receiving them.
Quality Assurance Bot (QA) – Version 2
ODJFS utilizes process automation technology to support county wide training and communication to improve accuracy of case processing. This bot was deployed in May 2021, and it was first used to review active SNAP cases that were already processed for recertification. The QA bot monitors potential inaccuracies in income details entered by county workers in Ohio Benefits. The bot then generates a detailed report which is shared with counties for review. Every month, the QA bot reviews about 6,000 cases for potential earned income issues.
Hours saved: This bot is estimated to save over 100 hours a month in QA operations.
The Baby Bot is used to add a newborn’s information to their mother’s Medicaid case. The Baby Bot corresponds with managed care plans, and with Ohio’s 88 counties when it encounters exceptions while processing. This is done to reduce turnaround time for processing. In March 2021, the Baby Bot was launched. Since the deployment of the bot, Ohio’s ability to provide newborns with care skyrocketed. Before the deployment of the Baby Bot, it would take counties seven to ten days to process Medicaid services for babies. Since deployment, the bot has provided well over 50,000 newborns with access to Medicaid within the same day of receiving information.
Since deployment, the Baby Bot has provided well over 50,000 newborns with access to Medicaid within the same day of receiving information.
The MyCare bot processes MyCare waiver flips in Ohio Benefits. This is done to ensure that service costs for MyCare are properly met. MyCare is a coordinated effort to provide health care and long-term services and support for individuals who are eligible for Medicaid and Medicare services. Before the deployment of the bot, delayed MyCare processing resulted in cluttered cases and inaccurate capitation fees. Since the deployment of the MyCare Bot in July 2020, the stress of manual data entry for Case Management Agencies (CMA) and PASSPORT Administrative Agencies (PAA) has been reduced.
Hours saved: The MyCare bot has reallocated approximately 500 operation hours for workers and processed well over 6,000 cases since its deployment.
The deployment of these bots provides administrative assistance to challenges facing Medicaid, SNAP, and other benefit programs. Leveraging AI technology reduces time spent on repetitive tasks and stress from caseload burden. Below is a diagram of all the deployed bots, as well as proof of concept bots.
Source: The Ohio Benefits Program is “Bot” In (nascio.org)
AI is a complement, not replacement
Each of these bots has a specific function in the allocation and management of benefits to Ohioans. For more details on how each bot works, check out this report by the Ohio Department of Administrative Services. Many of these bots assist with simple administrative tasks that save time, allowing case workers to focus on other administrative duties. AI tools such as these may help speed up basic operations but do not replace the need for in person communication, support, and guidance navigating public service programs.