How a $5 mistake sparked AI transformation at the front desk

A health system’s front desk shapes its patient experience. Yet frequent interruptions, staffing shortages, and high-volume manual tasks make errors hard to avoid. 

For Dr. Todd Beardman at NKC Health, a $5 billing mistake became the motivation to take pressure off the frontline – and champion AI transformation. 

The hospital front desk: small errors with big opportunities

Dr. Beardman is an Internal Medicine and Pediatric Specialist as well as Chief Medical Informatics Officer at NKC Health. He knows firsthand how each interaction adds up, from insurance eligibility to medical claims. 

“When I go to see my provider, all of these micro emotions translate into my patient experience. Did I get a good parking spot? Did I have to wait very long in the waiting room?” 

A routine primary care visit for Dr. Beardman resulted in a $5 bill weeks later. He paid the copay on his insurance card, but a recent copay change resulted in a billing error. When Dr. Beardman looked into the issue further, he discovered a multi-step workflow that employees did not have time to update or maintain, causing the copay discrepancy. 

This small data-entry error exposed a much larger need to address a long-standing operational workflow, prompting Dr. Beardman to take steps to prevent future mistakes. 

Testing AI in front desk workflows

Armed with EMR eligibility data, Dr. Beardman wanted to test whether AI could accurately determine patients' copays to help the organization avoid costly billing processes.

“The question I ask is: how much does it cost to actually collect $5?” 

NKC Health began automating a previously manual workflow by applying clinical and business logic, but found the initial accuracy to be about 65%. 

Zero or blank copays often reflected unaccounted scenarios:

• Medicare vs Medicare Advantage
• Missing group numbers
• Small or out-of-network payers

Many failures weren’t AI errors, though. Rather, missing or inconsistent data caused breakdowns in the new workflow. AI then used a reference table to immediately exclude visits that should never charge a copay, such as nurse- or lab-only. 

Once these were modeled correctly, AI accuracy improved. The process also established human review for errors occurring when multiple copays could apply. 

Dr. Beardman gained confidence in the new AI workflow, and NKC Health realized automation corrects what manual processes often miss.

Automation outperforms manual copay collection

Looking long-term, NKC Health faced a choice: retrain front-desk staff to manually capture insurance data such as group numbers, or automate the process using scanned insurance cards. Front desk staff already had full workloads; the organization chose automation. 

When group numbers are missing, AI Agents complete three steps: 

1. Check for Medicare Advantage
2. Determine if an insurance card is on file 
3. Insert data into the eligibility request

The AI automation process also revealed that copay collection wasn’t the only issue. Many patients pay 20–30% coinsurance. In their cases, eligibility showed no copay because insurance details had changed after the visit, even when legacy data fields still contained values.

Automation successfully reduced the number of billing statements with trivial amounts, relieving NKC Health front desk staff of hours of manual work.

The takeaway: “Perfection isn’t the goal – being meaningfully better than manual processes is,” says Dr. Beardman.

AI relieves – not replaces – jobs 

The fear around job loss is real, and according to Dr. Beardman, it’s worth acknowledging. 

“Especially in healthcare, there’s no shortage of meaningful work. AI simply takes on the tasks people shouldn’t have to do, so they can focus on what they’re best at.”

More than 70 clinic positions at NKC Health remained chronically unfilled for 5 years or more. “We can never hire enough. We're using tools like AI just to tread water.” 

Dr. Beardman adds that even if the hospital could hire someone to do manual, repetitive work, it’s not work anyone can sustain or do well. That’s why NKC Health is leaning on AI.

AI requires a flexible mindset

There’s still work ahead, but an iterative AI approach is important. Dr. Beardman’s advice: Fail early, learn quickly, and adjust the workflow based on the data. “Before AI, this copay initiative would have taken 6 months to resolve. Using Flow Builder, we had it working in a day or two.”

Moving from an idea to an AI flow starts with identifying the right use case. The real challenge isn’t deciding what to automate, but knowing where the true bottleneck is. 

Manual processes can be accurate; however, they’re highly inefficient. “Five years is the last time somebody updated the copay on that patient before their appointment,” adds Dr. Beardman. 

“The whole point is we actually have the right information. I can see it in the application, but I can't do anything with it.” When staff have to enter financial fields at every encounter, this added pressure limits their abilities to administer the best possible experience. 

Where AI fits in the health system revenue cycle

AI task mining, data mining, and AI Agents are a starting point to eliminate manual effort altogether. Dr. Beardman shares five specific areas where AI improves the revenue cycle:

1. Provider variances in eligibility files 
2. Documentation discrepancies in national versus local telehealth providers
3. Classification errors in freestanding urgent care centers
4. Insurance denial process improvement
5. Referral notifications to close care gaps

AI transitions organizations towards intelligent operations

AI provides the opportunity to outgrow old data processes and their limitations. Health systems must embrace innovation where it’s needed most, so small inefficiencies don’t scale into ongoing, frustrating experiences for patients and staff.

Dr. Beardman’s experience shows how a $5 mistake kick-started an AI journey that’s become irreplaceable in improving more than just the patient experience. By solving targeted pain points with automation, NKC Health is moving from reactive workarounds toward smarter, more resilient hospital operations.

‍