Notable is using AI agents to radically improve the efficiency and effectiveness of healthcare workflows. The company’s AI agents are currently automating over a million workflows each day, with a focus on streamlining tasks that traditionally require significant manual effort. In a recent interview, Notable’s Chief Medical Officer, Dr. Aaron Neinstein, explained that the company’s AI agents are able to operate autonomously, taking action on their own without the need for human intervention. This is in contrast to traditional software, which typically requires manual input from human users.
One key area where Notable is seeing significant success is in the automation of administrative tasks. Dr. Neinstein highlighted the potential for AI agents to take on tasks such as insurance verification, prior authorization, and appointment scheduling. By automating these processes, healthcare organizations can free up their staff to focus on more high-value work. In addition to reducing the administrative burden on providers, Notable’s AI agents are also helping to improve the patient experience by reducing wait times and streamlining the care process.
Watch the full interview to learn more.
Rebecca Willumson:
Hi there. I'm Rebecca Willumson. I'm the publisher of Fierce Healthcare, and I'm here today with Dr. Aaron Neinstein, Chief Medical Officer at Notable. Dr. Neinstein, thank you so much for joining me.
Aaron Neinstein:
Thanks, Rebecca. Great to be here.
Rebecca Willumson:
Now, AI is rapidly reshaping the future of healthcare, with AI agents emerging as a transformative force in improving efficiency, reducing administrative burden, and enhancing patient care. But what exactly is an AI agent, and what sets Notable's AI-driven approach apart? Today, Dr. Neinstein is here to help us break it all down. We'll explore why AI is gaining momentum, the impact it's already having on healthcare operations, and what the future holds for this technology. So to start us off, Dr. Neinstein, AI agents are being mentioned everywhere. Tell me, what exactly is an AI agent?
Aaron Neinstein:
If you think about traditional software that we use, like an electronic health record system, they wait for a human to enter data or trigger a prompt in action — and so traditional software is passive. It's storing information, like patient data. It lets a human log in and perform tasks. But as a user, you have to tell it what to do step-by-step. For example, if you think about what your clinic staff do before a visit, they usually have to manually verify information. They have to go through and scrub the chart, looking for health maintenance topics like cancer screenings to surface for the physician. Physicians and other providers have to spend their time manually documenting notes and manually verifying billing codes.
If you think about the EHR that we interact with today, it's essential for our workflows and for data storage, but it still requires a ton of manual input and effort from doctors, nurses, coders, and administrative staff.
What this means for the average patient is they're stuck waiting. There are barriers and friction that they experience in their care and in their access to care because of all the manual work required.
If you look at the average hospital health system, a typical organization has thousands of these tasks stuck in EHR work queues. We looked at an organization recently with 7,000 active work queues in the EHR. Every work queue is a task list, and every row is a patient waiting for their prior auth to be done, for their referral to be scheduled, or for their billing to be completed.
Let’s contrast that to what AI agents do. They work differently — they’re autonomous, intelligent software that can actually act on their own. In place of manual work triggered by a human, it's proactively completing tasks. It looks at information coming in the electronic health record or other systems, and acts on it automatically.
So let’s revisit the example I gave before of a human reviewing the chart to identify cancer screenings that are needed. Imagine, before the visit, the AI agent actually going in and scanning through the chart, looking to see if a mammogram has been done, or if a colonoscopy has been done, and then taking the next action to get authorization for that test or procedure to be done, to then go ahead and schedule it directly with the patient without a human person needing to take action. AI agents are proactive, and the best part is they continue to use the same underlying system. So your staff are still using the electronic health record, the workflows remain the same, but you have this extra digital workforce, essentially, that are now taking action on behalf of your staff, who are already very taxed in today's health system.
Rebecca Willumson:
Now, lots of companies seem to be talking about similar topics. Can you tell me what makes Notable's AI unique?
Aaron Neinstein:
So there are a few different things. First, we see a lot of large tech companies that work across industries developing AI agents. Those are companies that work also in financial services and travel and automotive. At Notable, we've been focused only on healthcare operations and clinical healthcare for about a decade, so all of our AI algorithms, all of the building blocks that we use to create our agents are completely focused on healthcare. We're deployed across more than 12,000 sites of care in more than 40 states across the US, so we have very healthcare-specific DNA.
The second thing, which is connected to that, is across my career, I worked in health IT at UC San Francisco for more than a decade. The thing that often separates companies that succeed from those that fail is actually the integrations. You can make really cool software, but if it's not tightly integrated into the data infrastructure of the core systems that a health system is using, like the EHR, and if it's not integrated into the workflows of the doctors, the nurses, and the staff, they fail. Notable has been building the integrations with workflows and data systems for the better part of a decade.
The third thing is actually implementations. Again, it's not just the software product that makes something succeed. It's the integrations and it's the implementation and delivery methodology. We’ve developed, over the past decade, our best practices for implementation for deployment, and we package those up in a customer success program we call Peak, which is all about helping organizations learn from each other and implement AI agents successfully.
If you take all of those things together, the healthcare-specific expertise, the integrations with health system core systems, and the implementation in customer success experience, it means we're able to deliver both an out-of-the-box library of AI agents that organizations can turn on and get benefit from right away, and at the same time, with our Flow Builder platform, it allows organizations to build their own custom AI agents. That combination of out-of-the-box impacts where we know you can get value quickly, and then continuing to work with Notable to configure and develop your own AI agents over time — I think all of those things make us unique and differentiated.
Rebecca Willumson:
So tell me, why do you believe the time is right for AI agents?
Aaron Neinstein:
If you look at the macro trends we're facing in the country, we're seeing one in five Americans are going to be of Medicare age by 2030. As they age into Medicare, chronic conditions are increasing in prevalence. At the same time, we have an epidemic of burnout among our care teams and our providers. I talked to organizations recently where they're expecting more than 25% of their primary care workforce to be gone over the next few years.
And as we talked about earlier, our care teams, our providers, and our staff are already struggling to keep up with the needs of patient care with current panel sizes and current staffing ratios. We already require five to 10 administrative staff per provider in the US. But as these trends continue and we have more care needs and fewer providers, there's simply not going to be enough staff available to deliver all of the care that's needed across the US. So in my mind, we've sort of passed the point of “should we use AI in healthcare?” To me, we're at the point of “how can we afford not to?”
If you step back and think about the type of care that we all want, imagine if you walk up to a front desk or if you're managing a clinic and you don't just have a few staff who are doing everything and they're underwater. Imagine having an unlimited number — 50, 100, 150 — of your best front desk staff. Every time you need a prior auth, it's done. Every time you want to schedule an appointment, you can. Every time the pharmacy faxes a refill request, it's taken care of quickly. Imagine, in the days leading up to a surgery, if every patient gets a phone call every day from a caring, attentive caregiver, screening them, asking the right questions, giving them the right education and information, taking as much time as they need to answer those questions. And the same thing after the surgery, when they leave the hospital. I mean, we think of it as concierge-level care. It feels impossible and it would feel like it's too expensive to be able to deliver that level of care.
But what if we actually can do it? And the answer isn't infinitely scaling up humans across the healthcare system that we can't find and we can't afford, but actually leveraging AI agents to deliver greater care and a more concierge-level experience to everyone.
Rebecca Willumson:
Now, Notable's AI agents automate over a million workflows daily. What is the opportunity for AI agents to transform healthcare productivity? And what operational workflows are benefiting from AI agents?
Aaron Neinstein:
Oh, gosh. So many. Let’s pick a couple of examples; let's talk about value-based care. To do value-based care well today, the assumption is that it takes an immense amount of human labor. Organizations that are succeeding in value-based care have large numbers of staff doing manual chart reviews, scrubbing the charts before every visit to identify care gaps, cancer screenings that are open, closed, completed, or not. They've got large numbers of staff making phone calls to patients to bring them in for appointments. They've got large numbers of staff making sure that the charts are coded properly so that they understand the risk level and the comorbidity levels in their patient population. They've got large numbers of staff trying to identify patients and enroll them in care management programs like diabetes prevention programs or other similar chronic disease management programs. They've got large numbers of staff doing care coordination after a hospital discharge.
That's a lot of human labor. And again, we have two million open administrative job requisitions today across the United States that are unfilled, so most places can't afford the labor. If they find the labor, there's a 30% turnover rate. So it's very hard to both afford and keep these positions filled. This is one example of the opportunity to, again, plug in AI agents to pick up and perform many of those tasks on behalf of humans today.
If you think about those impacts, at some of our customers who are using Notable for value-based care, we've seen thousands of additional care gaps closed, and thousands of additional diagnostic codes identified and picked up to help improve the risk adjustment and the understanding of the population. This is work that there just simply aren't enough humans around to do today.
Rebecca Willumson:
So tell me, what's one example of a story where you've seen an AI agent make a meaningful difference?
Aaron Neinstein:
Think about the last time you or a family member had to wait for a prior authorization. This could have been a specialty medication, someone was waiting for a CT scan or MRI, or even a surgical procedure. Today, it's a black box process; it’s anxiety, waiting, and feeling like you're fighting the system.
We actually had an example recently where, from the time the doctor placed the order in the EHR for an MRI, to where the prior authorization was completed, it was under 20 minutes. So that means you're sitting in the office with your doctor, they order the MRI, and rather than going home and feeling like you've got a week or two of fighting and worry and wonder, it wasn't weeks, it wasn't days, it wasn't hours. It was 20 minutes for the MRI authorization to be completed. And that's because there was an AI agent working in the background, automatically reaching out to the insurance company to get that done while the doctor and patient were still sitting there completing the visit.
Another example I'm really proud of is that patients have a similar experience when it comes to getting referrals done. When you get referred to a specialist, it often feels like it goes into a black box. Did the cardiologist get that referral? Are they actually going to respond to it? What's the status? Has anyone even picked that document up off the fax machine? And it turns out, still, millions and millions of referrals across the country are still done by fax, meaning that healthcare organizations today are paying an army of human labor. Again, this is 20, 30, 40 people at average in large health systems literally taking fax documents and typing information off the faxes into the electronic health record. That introduces delays in patient care. While that document is sitting on the fax machine, the patient is waiting to find out if their referral has been received well.
Well, we've automated that process of taking the fax document and uploading the information into the electronic health records, so now it's immediately available. At an organization where we did this, we went from an average turnaround time from referral received to scheduling the appointment of 14 days before deploying this solution, to two days.
I think both of those are examples with referrals and with prior authorizations, where the average patient today — anybody you run into at a cocktail party — they're going to have stories of frustration of dealing with those processes. We’re making them much more streamlined, much faster, with the use of AI agent technology.
Rebecca Willumson:
Well, it's clear that AI agents aren't just a glimpse of the future. They're here now. They're driving meaningful change across healthcare, improving operational efficiency, and delivering a better patient experience. The potential really is enormous. Dr. Neinstein, thank you so much for joining me today and for sharing your insights on how Notable's AI platform and intelligent agents are setting a new standard in healthcare automation. We'll be watching closely as this technology continues to evolve.