AR Glasses Replace Months of Job Coaching in 15 Minutes for Workers with Disabilities

Florida Atlantic University

Only about 15% of people with intellectual and developmental disabilities (IDD) hold competitive, integrated jobs. That number has barely budged in decades, despite federal programs designed to help. The bottleneck is not a lack of willingness or capability. It is a lack of consistent, scalable support. Job coaches burn out and turn over. Training timelines stretch for months. And the complex, real-world tasks that many jobs require, reading, listening, sequencing, remain stubbornly difficult to teach through conventional one-on-one methods.

A team at Florida Atlantic University decided to test whether augmented reality could do the job coach's work, literally. Their results, published in Focus on Autism and Other Developmental Disabilities, suggest the answer is a cautious but striking yes.

From 14% accuracy to 93% in a single session

The study placed participants with IDD in a real job: shelving books as library assistants. This was not a simplified task chosen for easy wins. Shelving books correctly demands reading call numbers, applying alphabetical and numerical sorting rules, and physically navigating stacks. Researchers specifically selected it because it requires the kind of multi-step cognitive work that people with IDD are often assumed to struggle with.

During a baseline phase with no AR support, participants completed an average of just 14% of task steps correctly. The job was genuinely hard for them.

Then the AR application was introduced. It delivered real-time, context-specific guidance through the device as participants worked, breaking complex sequences into manageable prompts that appeared at the right moment. Performance jumped immediately. Average accuracy climbed to 93%, with some participants hitting 100%. All met the study's mastery threshold: at least 90% accuracy across four consecutive sessions, completed independently.

Perhaps the most remarkable number: participants reached 75% accuracy and independence after just a 15-minute AR-supported training session. Traditional job coaching typically requires two to four months to achieve comparable results.

How the AR coach works on the job floor

The AR application functions as a portable, always-available job coach. Rather than relying on a human trainer who may or may not be present, the system uses the device's sensors to understand the worker's current context and deliver step-by-step visual and audio prompts. If a worker is holding a book and standing in front of the wrong shelf section, the system can redirect them before the error is made.

This approach addresses two persistent problems in supported employment. First, the chronic shortage of qualified job coaches. High turnover in these positions means workers frequently lose their support person and must start over with someone new. Second, the scalability problem: a human coach can only work with one person at a time, while an AR system can theoretically support many workers simultaneously with minimal added cost after the initial development investment.

Ayse Torres, the study's senior author and an associate professor at FAU's College of Education and College of Engineering, framed the potential in terms of system-wide efficiency. If the technology can help people work more independently while stretching program resources further, it creates benefits for workers, employers, and service providers simultaneously.

The supported employment gap the technology could fill

Supported employment programs have existed since the 1980s, backed by federal funding and a clear philosophical commitment to integrated work for people with disabilities. But the programs have never fully delivered on their promise. The 15% employment rate for people with IDD reflects a system that struggles with consistency, scale, and the sheer complexity of preparing someone for a job that changes day to day.

Job coaches are the backbone of these programs, providing on-site training, social support, and troubleshooting. But the role is demanding, often low-paid, and characterized by high turnover. When a coach leaves, the worker they supported may regress, and the cycle starts again. Meanwhile, the social and environmental challenges of a workplace, navigating coworker interactions, adapting to schedule changes, handling unexpected situations, persist regardless of whether a coach is present.

AR does not solve all of these problems. It cannot replace the social and emotional support a good job coach provides. But it can address the instructional component with a consistency and availability that human staffing models have never achieved.

Small sample, controlled setting, early days

The study's limitations are worth stating plainly. This was a small-sample, single-case experimental design, the standard methodology for early-stage assistive technology research but not a large-scale trial. The participants worked on one specific task in one specific setting. Whether the AR system would perform as well across different jobs, workplaces, and populations of workers with IDD remains an open question.

The study also took place in a somewhat controlled environment. A library is structured, predictable, and relatively quiet compared to a busy restaurant kitchen or a retail floor during holiday season. The degree to which AR coaching can handle the noise, unpredictability, and social complexity of other work settings has not been tested.

There is also the question of long-term retention. The study demonstrated rapid skill acquisition, but whether workers maintain their performance over weeks and months without continued AR support is not yet known. And the technology itself, while increasingly affordable, still requires hardware, software maintenance, and initial setup that may be beyond the capacity of some smaller supported employment programs.

What comes next for AR-assisted employment

The research team, which included faculty from FAU's departments of Special Education, Biomedical Engineering, and Electrical Engineering, designed the study as a proof of concept. The next steps would logically include testing with larger groups, across multiple job types, and in less controlled work environments. Cost-effectiveness studies comparing AR coaching to traditional models would also help determine whether the technology makes financial sense at scale.

The broader context matters here. Technology is reshaping the workforce rapidly, and people with disabilities risk being left further behind if accessible tools do not keep pace. AR job coaching represents one approach to closing that gap, not by replacing human support entirely, but by making the instructional component of job training more consistent, more available, and dramatically faster.

For the workers in this study, the difference was stark: from failing at a task to mastering it, not over months of coaching, but in a single afternoon.