5 Q’s for Malinka Walaliyadde, Co-Founder of AKASA – Center for Data Innovation
The Center for Data Innovation spoke with Malinka Walaliyadde, CEO and co-founder of AKASA, a company that provides AI-driven solutions for healthcare revenue cycle management. Walaliyadde discussed how the company’s platform leverages generative AI to streamline billing processes, improve medical coding accuracy, and help healthcare providers optimize financial operations while enhancing the overall patient experience.
Martin Makaryan: Can you start by introducing yourself and sharing the inspiration behind the founding of AKASA?
Malinka Walaliyadde: At a high level, AKASA uses generative AI to help hospitals and healthcare providers manage their billing and insurance claims—a process known as revenue cycle operations. This ensures that providers get paid accurately and efficiently for the care they deliver. By streamlining this system, AKASA helps reduce administrative burdens, improve hospital efficiency, and enhance the patient experience.
We founded AKASA because, while the quality of American medicine is among the best in the world, the healthcare system itself is overly complex and inefficient. Over time, billing and payment processes have become increasingly complicated, creating frustration for both providers and patients. With advancements in generative AI, we now have the tools to simplify these processes—making them faster, more accurate, and easier to navigate.
Makaryan: Why is healthcare revenue cycle management such a challenge, and what makes it a good area for AI-driven innovation?
Walaliyadde: The biggest challenge in healthcare billing is how medical records connect to the revenue cycle—the process hospitals and clinics use to bill insurance companies for patient care. Since payments are based on the treatment a patient receives, billing teams need to interpret clinical records accurately.
Historically, this was a slow and error-prone process because the staff handling billing aren’t medically trained. They have to manually sift through complex medical records to determine the correct billing information, leading to inefficiencies and potential mistakes.
With AI, we’ve trained a large language model that understands both clinical records and financial data. This AI serves as a co-pilot for billing teams, helping them quickly and accurately extract the right information. By automating this process, we increase efficiency, reduce administrative burdens, and improve overall financial operations in healthcare.
Makaryan: How does AKASA’s AI-powered platform work, and did you build a proprietary model or fine-tune an existing one?
Walaliyadde: At the core of our platform, we use open-source large language models as a foundation. Think of this as starting with a model that has general knowledge—like a high school graduate—but lacks specialized expertise. We then pre-train and fine-tune it using healthcare-specific financial and clinical data, transforming it into an expert in healthcare revenue cycle management.
The result is a model that outperforms industry-standard models like GPT-4 by 40 to 50 percent in our domain. Beyond that, we further adapt the AI system for each hospital or healthcare provider we work with, training it on their specific data and documentation practices. This ensures the AI system understands the unique nuances of each organization and can provide tailored, accurate support to billing and revenue cycle teams.
Protecting patient data is our top priority, and we follow industry-best security standards to maintain trust with healthcare providers. We retrieve data through standard healthcare integrations that allow our AI to learn from real-world examples without compromising patient privacy. Our approach is designed to be seamless for providers, minimizing technical burdens while ensuring compliance with regulations.
Makaryan: Can you share any success stories or tangible impacts AKASA has had on healthcare organizations?
Walaliyadde: One of the most tangible ways AKASA has helped healthcare organizations is by improving medical coding—the process of translating patient diagnoses, treatments, and procedures into standardized codes used for insurance billing. Accurate coding ensures that hospitals and clinics get properly reimbursed for the care they provide. However, trained coders are becoming harder to find, leading to delays, inefficiencies, and lost revenue.
Our AI system supports coders by reviewing their work and suggesting improvements, increasing both accuracy and efficiency. The financial impact has been significant, with some healthcare providers recovering millions in revenue they would have otherwise missed due to coding errors. Just as importantly, coders find the AI easy to use and helpful, improving their overall experience, a key measure of success for us.
Makaryan: What is your vision for the future of AKASA and AI in healthcare?
Walaliyadde: We see AKASA as an AI partner that reduces administrative burdens and optimizes revenue cycle management, allowing healthcare staff to focus more on patient care rather than paperwork. As AI continues to evolve, its role in simplifying healthcare operations will only expand, leading to better financial outcomes for providers and an improved experience for patients.
One of our biggest focuses is ensuring that every product we develop delivers real value to customers. We continuously refine our solutions by working closely with healthcare providers, gathering feedback, and iterating based on their needs. Another key priority has been clearly communicating the benefits of AI-driven automation and ensuring a smooth adoption process within health systems.
