Full Transcript:
‍

Shruti Mehrotra: Hi, welcome everybody. I'm Shruti Mehrotra, Head of marketing at Aarogram. We are going to kickstart the event now. I would like to say that we're meeting in interesting times. Healthcare continues to face mounting challenges in 2025 with staffing shortages, rising patient demands, and administrative bottlenecks.

In the midst of it all, AI is emerging as the game-changer we've all been waiting for. With predictions that AI will soon handle up to 60% of front office tasks in healthcare practices, the future is not just coming, it's here. From automating repetitive front office tasks to tackling complex decision-making, the possibilities are endless. And today we are diving into how this technology is transforming operations in real-world settings.

I'm thrilled to introduce our speakers. We have Oscar Ortega, Managing Partner at Cream Digital, who has been instrumental in setting up AI infrastructures that streamline business operations at a range of businesses including medical offices. And we also have Kashyap Purani with us who is the brains behind the revenue operation automations at Aarogram. Together they unpack how AI can change the face of front office operations, freeing up time for staff and enhancing patient satisfaction. Let's dive in. The floor is yours, Oscar.

‍

Oscar Ortega: Thank you so much. I really appreciate your time and Kashyap, I appreciate you taking the time with us so we can talk about everything AI. Let me just get right to it.

‍

Kashyap Purani: Yeah, I'm very excited.

‍

Oscar Ortega: What you see here is an AI infrastructure that we've been able to build out for different medical practices. And to Shruti's point, getting into you know, kind of some of the pain points in in the medical space right now, the way that we look at medical practices and how we can solve their problems in the front desk is we look at ways or the different tasks that each respective position handles, right? And so let's say take for example the AI patient coordinator. Here we look for the key tasks that they handle. For example, they handle call routing, call handling, collecting patient information, verifying insurance, collecting payment, managing paperwork, data entry, etc., right?

Then we look for ways to leverage AI to either optimize, automate, or eliminate that role that they must do entirely so they can focus on higher ROI tasks, right? And what can I take off their plate immediately would be patient information collection, insurance verification (now we can do that with Aarogram), automated scheduling, document management, data entry, and billing, right? And so those are some of the key tasks that today we can immediately handle or hand over to AI in a way that is effective, optimized, and again brings a nice ROI to medical practices.

So, first things first, the first thing we're going to talk about on the front desk infrastructure is the what I call the intelligent call routing, or automated call routing. And so, this is especially crucial for medical practices that have a high call volume. And so, oftentimes they have a tonne of missed calls from patients. They can't get to them all. And not every single phone call is of the utmost importance, right? And I mean that respectfully. Some of these phone calls from patients might just be seeking information on your practice. Hey, what are your hours? Do accept this insurance provider, what are your services, what are your provider's name, right? And what's your address? Those questions can be handled with AI very easily, right? I mean AI can handle these things 24/7, 365 without breaking much of a sweat. And it allows your staff to handle the questions or the or the calls that are kind of more related to again higher ROI tasks, right? So. first things first, we install an automated call system to route these calls intelligently to and from departments. And then we also have the intake portion, right, where we're able to automate the practice's workflow.

Again, it can route calls to and from different departments or specific employees in a way that feels very natural. For instance, if the practice is located on Southwest A Street, it will recognize the location and transfer the call accordingly. It doesn’t have to function like a static and stiff IVR system. This approach is dynamic and adaptable.

It can also ask qualifying questions. For example, if a patient wants to book an appointment, it can determine whether they are a new or existing patient. Based on the parameters set, the call is directed appropriately. You may prefer the patient to book an appointment through AI or directly with your patient coordinator. The first step is understanding this intelligent call routing, which serves as the initial layer of defense.

The next layer involves patient intake, which can be done in various ways: via SMS, WhatsApp, email, your website, or AI voice. The goal is to make it as easy as possible for patients to book appointments with your practice in a manner that is both accurate and secure. We can integrate with 99.99% of all EHR or PMS systems. If an API isn’t available, we have compliant and safe workarounds. The objective is to sync your calendar to prevent overbooking or double booking while asking qualifying questions to ensure patients are tagged with the correct visit type. This information is then added into the EHR or staff is notified of new appointments.

The entire process is streamlined and can be executed using AI in English, Spanish, or other languages based on regional demographics. For example, we have a client in New York with many Russian patients, so we built an AI bot that operates in both English and Russian to intake those patients effectively. This system is versatile and ensures accurate onboarding while confirming patient information.

One particularly effective method for insurance verification involves sending patients a link where they can upload their insurance information. They can take pictures of their insurance card (front and back), and OCR technology extracts the details for verification. This leads to the next step: insurance verification. Aerogram’s smart verified tech offers significant advantages by not only verifying if insurance is active but also determining what the patient will owe for their visit type at the CPT level. This ensures clarity about coverage and costs before the visit.

One major pain point in medical practices is that patients often don’t know how much they will owe due to delays in insurance verification—sometimes taking hours, days, or weeks. Additionally, prior authorization may be required for certain services. With smart verified tech, these issues are flagged early so practices know which patients need authorization and can expedite the process for faster service delivery and higher patient satisfaction.

Another common issue contributing to no-shows is uncertainty about costs. Patients may hesitate to attend appointments if they’re unsure whether bills will be $50, $75, $200, or more depending on the visit type. Providing clarity about costs upfront allows practices to notify patients ahead of time, set up payment plans for higher bills if needed, or collect payments before appointments. This reduces no-shows and ensures revenue stability while eliminating surprises for both patients and staff.

This platform optimizes front desk workflows by streamlining processes from appointment booking to cost transparency. For example, a patient could call after hours—say 7 p.m. on a Friday—book an appointment, have their insurance verified, receive an email or SMS detailing their expected costs for that visit type, and even complete payment via an automated link before arriving at the practice.

This infrastructure addresses key challenges in medical practice management while improving efficiency and transparency. Statistics show that one out of two physicians report burnout due to administrative burdens, spending an average of 15.5 hours weekly on such tasks—eight of which are spent on EHR systems alone. By leveraging AI for these tasks, practices can optimize their front desk operations and free up valuable time for strategic growth decisions.

In summary, this front desk AI infrastructure helps streamline administrative workflows while enhancing patient satisfaction and revenue predictability. I’m happy to answer any questions you may have!

‍

Questions:

As a healthcare practitioner, administrator, or someone managing administrative or revenue operations in healthcare, how can I get started with automating processes using AI? Are there specific use cases or tasks that are easier to begin with, which could help in making both staff and patients more comfortable with the transition?

‍Answer: One of the first steps we suggest is addressing the front line of defense: answering all phone calls. A major complaint in the medical field is that calls often go unanswered, leaving patients on hold for extended periods—sometimes 40 to 50 minutes or longer. When call volumes are high, practices can either hire more in-house staff or outsource this process. However, outsourcing has limitations in terms of what the agents can say or do.

If you focus on handling the front desk effectively, there are several strategies to implement. For example, AI can transfer calls to specific employees or departments. If a department is busy and unable to answer within a set timeframe (e.g., 30 seconds to a minute), the system can redirect the call to another AI agent. This agent can take messages, ask qualifying questions, and route the information securely to the appropriate department or provider. In emergencies, the AI agent can immediately send an SMS to the doctor, ensuring they are informed about urgent situations.

The AI system can also gather additional details from callers through qualifying questions and relay this information in a safe and secure manner via SMS or email, depending on client preferences. For instance, it could notify a specific department with a message such as: "Hey Kashyap, we just received a call for this department regarding [specific issue]." This allows staff to promptly address patient concerns by contacting them directly.

By leveraging AI agents for front desk operations, practices can ensure that patient calls are answered efficiently, and critical information is shared securely. This approach minimizes delays and enhances communication between patients and medical professionals.

‍

How do you ensure that multiple tools used for automating different tasks in healthcare work together seamlessly while maintaining HIPAA compliance and ensuring that patient data remains safe and secure?

Answer: We can view it like a symphony, where every tool is an instrument that must work harmoniously together. For instance, we might use AWS EC2 to store information or as a database. To ensure compliance, we sign a Business Associate Agreement (BAA) with Amazon, ensuring all data is encrypted. We can run our automations from this platform in a HIPAA-compliant and secure manner.

We also utilize other platforms, such as Go High Level, which offers a HIPAA-compliant component that we can integrate into our workflow. The key is to use multiple HIPAA-compliant platforms and make them work together seamlessly as part of our infrastructure. This ensures that the entire process is smooth and efficient.

The first step we take with any new setup is to ensure that all tools are integrated in a way that maintains HIPAA compliance throughout the process.

Let me know if further refinements are needed!

‍

Is there anything different in modern AI solutions that makes integration with other Healthcare tech tools easier than before?

Answer: On the AI voice side, I'll tell you that it's just gotten better. Whereas about a year and a half ago it wasn't quite there yet and was still kind of rudimentary, now it's gotten to the point where it can handle Level 1 and Level 2 tasks. The ability to give it tools on the call—for example, to make an API call to the CRM or send information in real time—allows it to access databases, knowledge bases, and CRMs in real time. This enables it to answer questions in much more detail than it could a couple of years or even a couple of months ago.

This is the era or the day of the AI agent. Earlier this year, we talked about how 2025 was going to be the year of the AI agent. The idea behind building these systems is to provide tools that allow the AI to complete entire tasks without needing human intervention halfway through.

When you say AI agent, by 'agent' we mean like we automate the whole process end to end. You are kind of hands off where you are relying on AI to not just do it one part of somewhere you know it's like kind of doing its work as an agent that's why you're calling it agent.

I’ll give you a couple of examples of humans in the loop when an AI agent is being used. If we have a specific medical practice that deals with pediatrics, for instance, they may require a human in the loop if the patient is calling about a specific procedure. In such cases, they wouldn’t want AI to handle it. When we ask a question and receive an answer that triggers a workflow, the system can transfer the call to a human or the relevant department to handle it appropriately.

With AI, the goal is to build systems that automate repetitive tasks. As I tell my clients, if you already have repetitive workflows in place, AI can handle those for you. For example, not every EMR platform has APIs, so we create workarounds. We can build an AI agent hosted securely on a server to receive data and safely input it into the EMR system. This entire workflow—from receiving patient calls to transferring them to the intake department, asking questions, and adding data into the EHR system—can be done entirely by AI without requiring human intervention. However, if needed, staff can still be alerted about new appointments or errors during the process.

For instance, staff might receive a notification like: "Hey, Kashyap has booked an appointment for 2 PM on Wednesday." If there’s an error during this process, staff can review it and fix any issues within the workflow. The idea is to give AI autonomy while allowing room for human intervention when necessary.

When starting out with automation, we recommend beginning with answering phone calls and intaking new patients. These initial steps alone free up significant time for staff to focus on higher ROI tasks. Once these workflows are established, additional steps and upgrades can be added gradually.

‍

Can Aarogram's AI agent solutions be integrated to physical therapy EMR systems like web PT or prompt?

Answer: Yes, absolutely. The way we integrate automation is by building an AI agent using Robotic Process Automation (RPA). This allows us to fine-tune the AI agent to follow your specific workflows. For example, we can design it to handle patient intake based on your existing processes. You let us know what that process looks like, and we build the system accordingly. The AI agent can then capture data from various sources—such as phone calls, SMS, or other inputs—and accurately input that information into your Electronic Health Records (EHR) or Practice Management Systems (PMS). This ensures the entire process is automated and completed end-to-end.

In fact, we work with some physical therapy practices where customers use tools like WebPT and Prompt. We integrate with these systems seamlessly, and we’re also in direct contact with their teams for support. As you mentioned, with AI tools—whether it’s RPA or a next-generation AI agent—it’s now possible to work across different systems, even when using interfaces that weren’t originally designed for integration. This opens up a whole new world of interoperability and system integration in healthcare.

One of the key benefits of this approach is improving Return on Investment (ROI) for administrative staff by increasing productivity and efficiency. But what about patients? Are they ready to interact with these systems? Will this technology build trust with patients or reduce it? Transparency is key here. Sometimes patients don’t even realize they’re interacting with AI, but in certain cases, being upfront about using automation can actually enhance trust. Patients may appreciate knowing that their provider is leveraging advanced technology to be more efficient, responsive, and proactive.

As a patient myself, I would value a provider who can act immediately without unnecessary delays. Automation not only benefits staff but also improves the overall patient experience by reducing wait times and ensuring faster responses.

‍

Can you share examples of implementing AI agents in healthcare systems in the form of direct patient feedback?

Answer: After handling over 100,000 calls, the biggest insight I can share is this: you cannot mislead your patients. Transparency is crucial. I advise my clients to immediately inform patients that they are interacting with a virtual assistant. This approach builds trust by emphasizing that it’s the fastest way to book an appointment. One of the most common complaints in medical practices is how long it takes to complete tasks, so letting patients know upfront that the system can book their appointment within minutes establishes confidence.

We also use tools to enhance this trust. For example, the system can recognize a caller’s phone number and say, "Hey Kashyap, I see your phone number is 305-632-5362. Is that the best number to reach you at?" When the patient confirms, it creates a personalized experience that sets it apart from traditional automated systems. Ultimately, patients want their problems solved—whether it’s booking an appointment or answering a question. They don’t necessarily care if it’s AI or a human; they just want efficiency.

That said, there are about 3–5% of patients who prefer not to interact with AI and request immediate transfer to a human agent when they realize it’s AI. This is completely normal and easy to accommodate—the system can seamlessly transfer such calls to the appropriate department or staff member. For the majority of patients, however, AI’s conversational capabilities and ability to access tools in real time make it highly effective.

The goal is simple: provide solutions efficiently. Whether through AI or human intervention, patients value responsiveness and problem-solving above all else. By leveraging AI tools for operational automation in healthcare—such as virtual assistants, appointment scheduling systems, and patient engagement platforms—practices can streamline workflows while maintaining transparency and trust.