YouTube Deep SummaryYouTube Deep Summary

Star Extract content that makes a tangible impact on your life

Video thumbnail

they’re replacing insurance companies

Founders, Inc. • 2025-07-27 • 3:46 minutes • YouTube

🤖 AI-Generated Summary:

Revolutionizing Healthcare Billing: How AI Voice Agents Are Fighting Insurance Claim Denials

Healthcare billing in the United States is notoriously complex and frustrating, not only for patients but also for hospitals. A striking example lies with UCSF hospitals, which collectively generate approximately $536 million in revenue each month primarily through insurance claims. Yet, despite this massive volume of claims, UCSF reported a staggering $1.4 billion in uncompensated claims in 2023 alone — bills that went unpaid.

The Root of the Problem: Insurance Claim Denials

Each month, UCSF processes around 170,000 insurance claims, but about 20% (or 30,000 claims) get denied almost immediately. This means thousands of patients are suddenly responsible for paying large medical bills out of pocket. While many factors contribute to the challenges in American healthcare, claim denials stand out as one of the most fixable problems.

When a claim is denied, the hospital must actively fight the insurance company to get it approved. This process used to be straightforward—one phone call or document submission by a doctor was often enough. Today, however, the process is deliberately convoluted, requiring multiple phone calls, dozens of documents, and lengthy back-and-forth communications that average four months. By the time a claim is resolved, thousands of new patients have been seen, and the backlog of unpaid claims continues to grow.

Enter AI Voice Agents: Automating the Fight Against Denials

Recognizing the inefficiency and intentional hurdles in processing claims, a team has developed an innovative solution: AI-powered voice agents designed specifically to handle insurance calls on behalf of hospitals. These AI agents can mimic human conversation, ask relevant questions, and navigate complex conversations with insurance representatives to get claims approved faster.

Hospitals like UCSF are using these AI voice agents to automate the thousands of daily calls required to process insurance claims—freeing up staff to focus on patient care rather than paperwork. One UCSF hospital processes about 5,000 such calls daily, and the AI agents have been deployed at 23 hospitals and medical billing companies, collectively handling thousands of calls every day.

How Do These AI Agents Work?

The development team discovered that a typical insurance call involves about 500 possible questions, making it impossible for a simple chatbot to handle effectively. Through extensive trials—including a test with UCSF involving 20,000 calls—they refined their AI agents to sound natural and realistic.

Key findings from their research included:
- Incorporating realistic background noises like office chatter or street sounds.
- Adding natural speech patterns such as filler words ("ums," "buts") and pauses.
- Surprisingly, male AI voices had a 40% higher success rate than female voices when interacting with insurance brokers.

By the end of their trial, the failure rate of one of their AI voice agents was just 2%, demonstrating remarkable effectiveness.

Real Impact: From Trial to Widespread Adoption

Following the success of their trial, UCSF is set to integrate these AI voice agents into five more of its hospitals. The system has already helped UCSF hit record-high revenue numbers during the trial phase, proving that automating insurance call handling is not just a technological breakthrough but a financial game-changer.

The creators of this AI solution bring years of healthcare industry experience, having personally navigated the frustrating insurance claim process and even built one of the largest healthcare agent networks in Taiwan. Their mission is clear: to make healthcare more accessible by removing bureaucratic barriers that prevent hospitals from getting paid and patients from receiving affordable care.

Why This Matters

This breakthrough highlights an important truth: while many problems plague the American healthcare system, some are solvable with smart technology. Automating the cumbersome, often intentionally difficult process of insurance claim disputes through AI voice agents could drastically reduce uncompensated claims, lower financial stress on patients, and allow hospital staff to dedicate more time to what truly matters—providing quality care.

As this technology scales, it promises to transform the healthcare billing landscape, proving that innovation and compassion can go hand in hand to fix one of the most frustrating parts of healthcare.

In Conclusion

The fight against insurance claim denials is one of the few healthcare problems we can actually fix right now. AI voice agents are proving to be an effective and scalable solution, helping hospitals reclaim millions in revenue, reducing patient financial burdens, and streamlining one of the most opaque parts of healthcare. As more hospitals adopt this technology, we move closer to a healthcare system that works better for everyone.

📝 Transcript (121 entries):

These pins represent UCSF hospitals. Collectively, they bring in roughly $536 million in revenue every month, which is mostly made up of insurance claims. Patients walk in, receive care, and the hospital sends a bill to the insurance company. But in 2023, UCSF reported $1.4 billion in uncompensated claims, which means that the bill wasn't paid for. Normally, UCSF makes 170,000 claims every month. Roughly 20% of these claims get denied almost instantly. That's 30,000 people who now have to front the bill. Now, there are a lot of reasons why healthcare sucks in America. But fighting for claims is one of the worst problems we have. And unlike all the other reasons, we can actually fix this one. >> When a claim gets denied, it's actually on the hospital to go and fight the insurance company to get it covered. >> Before the 2000s, it used to be just one phone call or one document where a doctor proved that they provided care. But today, it seems like they've deliberately made it harder than it needs to be. Five phone calls, each taking half an hour. 10 documents for every appeal. All of this back and forth takes 4 months on average. By then, 160,000 new patients have come and gone. So, the backlog of uncovered claims just keeps piling up. >> All because they're forcing hospitals to go through 50 hoops that were never needed. But there's a team here, and they found a way to fight back. Yo, you guys busy? >> Yeah. We're building an army of voice agents for hospitals to automate their insurance phone calls. >> Hospitals take up way too much time to process insurance claims and it's obviously deliberate. >> All of the information that the insurance company asks for through this call can be easily accessed through a insurance portal or some software solution. >> It's just information that can be sent instantly if they wanted it, but instead they force you to make these phone calls. They have built entire armies to prevent you from getting through. So we are building our own. How would you say how would you? So over the past few years, AI voice agents have become more capable than ever. >> They sound just like us, can ask real questions that we would ask, but most importantly, they can direct conversations to get what we want. >> So one of the hospitals we work with does 5,000 calls per day to insurance companies to get these claims processed. And we provide thousands of AI voice agents to basically trick the insurance companies and go through all the hoops for them. So hospital staff can do things that actually matter and save their time. >> But why can't a hospital just hook up chatbt to their phone? In a typical insurance call, there are roughly 500 possible questions that can be asked. Their team learned this through a trial they ran with one of UCSF's hospitals where they processed an initial batch of 20,000 calls. They tested different voices, styles of conversation and questions. They filtered it down to four of the best people in their AI army and two important qualities. Realistic background noise like office sounds or street noise and realistic filler words like ums, buts, pauses. And they also discovered that male voices have a 40% higher success rate than female voices when talking to insurance brokers. >> Weird. >> By the end of their trial, the failure rate of one of their voice agents was only 2%. Which is why UCSF is trusting them to integrate into five of their hospitals next month after hitting a record high in revenue during their trial. >> Right now, they're deployed in 23 hospitals and medical billing companies, handling 5,000 calls every day, freeing up the staff's time for actual patient work. They're doing this for their friends, their family, ordinary people because they've seen countless stories just like this. >> All three of them have worked in the healthcare industry for years. So, used to make these calls himself, and David actually built one of the largest healthcare agents in Taiwan with 120,000 providers using it every day. >> They've experienced this problem firsthand, which is why they know how to fix it. What are you going to do for the world? >> I want to make healthcare more accessible. [Music] Hello.