Alexis is director of content at AIMed, with responsibility for the research, development and delivery of products across events, digital and publishing. A highly experienced events executive with a career focus on the intersection between healthcare and technology, he is also a school governor leading on teaching, learning, and quality of education.
Exponential AI COO Nikhil Mendhi talks transformation, innovation, and the implementation of artificial intelligence across healthcare payer and provider organizations.
You have built a career leading large scale transformations. What are the top 3 tips you would give to anyone tasked with driving change in their organization?
You know, large scale transformations have always been a challenge to pull off. By some estimates, less than 30% of all transformation exercises succeed. With the onset of digital, industries are witnessing dramatic changes, making transformation exercises even harder. However, in my experience, transformations that are designed to be people centric have seen great success. From my perspective, the top 3 drivers of such success are:
- First, you must focus on the business case. This has to be very objective and aim for solving a genuine problem that can be quantified
- Second, empowering people to become change leaders. In the new customer-centric digital models, transformation requires not only capability enhancements but also requires organizations to drive cultural and behavioural changes
- Third, technology needs to be in the hands of people. Newer technologies such as, AI, blockchain, etc. must be easily accessible and adaptable to empower your employees to work in new ways
Healthcare is perceived as slow to adopt and scale innovation. What could the sector learn from other large enterprises?
The perception is true to an extent. Aerospace, e-commerce and fintech have successfully demonstrated how scaling of cutting-edge technologies can positively impact overall end user experience while making services more secure, transparent, fast and affordable.
But healthcare innovation isn’t far behind. Innovations are now happening at a formidable speed, overcoming the inherent data, technology, and regulatory challenges.
While this is already happening, healthcare should learn from other industries to design, execute and scale more proactive consumer centric processes by leveraging exponential technologies such as AI.
How is AI being used in healthcare for payers right now?
Payers have been making investments into AI over the last several years to solve some of their most pressing issues. A recent Gartner report estimates that over 50% of payer organizations have their AI strategy in place. Widespread adoption has been seen in customer operations such as member acquisition, member enrolment, customer retention, and so on. There have also been pockets of AI experiments on the operations side too, such as claims processing, data mining, and clinical audits.
With positive ROIs from their AI and ML investments, we now expect payers to go mainstream with AI focused on MLR as well as ALR operations over the next 5 years. We believe that AI will accelerate the convergence of payers and providers to eliminate administrative waste and improve health outcomes.
Is there traction with AI for claims audit? How specifically is AI deployed in this solution area?
We are seeing great traction for our AI-based Claims Audit Solution. It enables payers to scale their audit capability from a sample-based audit to a 100% claims audit using AI algorithms. It also allows them to move from postpay audit to prepay audit and reduce revenue leakage by a factor of 3X.
More importantly, the learnings that are now available through audit can be applied to prepay processes such as adjudication and payment integrity to improve accuracy and processing efficiencies. It can also be applied to provider environments to cut back on administrative costs related to rebilling, rework, recovery and denial management.
We want to take this one step further and apply provider-specific findings in provider environments to improve the accuracy of claims that we billed by the providers. This will significantly reduce the administrative cost from providers and also enhance their cash collection
You mention AI being used for payment integrity. How effective is it?
An estimated ~5% percent of all health claims are inaccurately paid today, stemming from issues ranging from simple data and clerical errors to intentional fraud. Payment Integrity or Program Integrity actively focuses on minimizing the financial losses arising from fraud, waste, abuse, COB/third-party liability, subrogation, error/clinical editing and administrative overpayment.
Traditional “pay and chance” is not effective, especially with increasing complexity of payments, contracting and regulation. AI offers tremendous potential for identifying improper payments in real time and the opportunity to stop these leakages before payment.
We offer a Payment Integrity suite of solutions focused on improving payment accuracy, reducing waste and fraud abuse in the healthcare environment. Our Decision Agent-driven PI solutions continually learn from data, decisions and feedback to offer new insights on potential savings across the prepay and post-pay Payment Integrity Value Chain. We help clients predict savings and reduce a huge amount of administrative and clinical waste from the healthcare system.
Operationalizing and scaling of AI is a universal challenge in healthcare. How does your organization approach this?
That’s very true. While AI promises great results, it’s quite a struggle to get there. Besides the challenges with solution design, model orchestration, training and productionizing, organizations are also often lacking sufficient clarity on speed to market, expected value and time-to-value.
By some estimates less than 10% of all the AI models built make it to production, making operationalization of AI solutions one of the bigger challenges.
We offer a Healthcare Enterprise AI platform, Enso, that simplifies the development, orchestration, deployment and management of AI solutions applied within complex processes. The platform comes with pre-trained healthcare assets and is designed for collaborations between the multiple personas required for development. In addition, superior orchestration capabilities and the engineering around the AI models enables a much more efficient way to deploy AI solutions and provides for continuous monitoring and management of AI and ML models in production.
The platform is designed to be the underlying foundation for all AI in the enterprise. For example, one of our payers uses the platform to manage over 30 AI solutions having around 900 AI and ML models. They have been successful in taking many solutions to production by switching a platform approach for their AI implementations.
Gartner has been talking about decision intelligence. Is this the way forward for really moving AI solutions into production?
In most enterprises, AI is primarily used as a predictive tool, which seriously restricts the potential that it has to offer. AI used for decision making can totally change how healthcare processes are currently run to autonomous processes delivering better cost and quality outcomes at scale.
While today AI is witnessing accelerated adoption within healthcare, using it for decision making still remains a challenge. For AI to be used for decision making, it needs to be used in collaboration with rules, optimization and other decision making techniques. This is known as Decision Intelligence.
Using AI for Decision Intelligence will lead to AI delivering transformative results and drive superior ROIs for the enterprise. The coming years will see AI as Decision Intelligence as the largest application of AI, not only in healthcare but also across other industries.
However, building, training and managing these decision pipelines is complex, and operationalizing is even harder. This stems from the limited capabilities in orchestrating complex systems, a lack of the necessary decision infrastructure, and complexities in monitoring and maintaining these systems.
To solve these challenges we enable a unique Decision Agent Approach on our Platform. Agents are intelligent digital workers, trained to take human-like decisions for a particular domain. Enso comes will multiple pre-trained decision agents that can easily be orchestrated, along with custom-built agents to build end-to-end AI solutions that deliver augmented and autonomous decision-enabled processes for enterprises.
What do you see as the greatest opportunities for AI in healthcare?
We are excited about the overall direction in which healthcare is changing. There’s a shift towards proactive, preventative approaches along with the increasing share and acceptance of digital care delivery.
AI solutions are going to be crucial to these efforts. Similarly, as personalized treatments find wider adoption, health systems will need to scale decision-making across administrative and clinical functions. AI will play a critical role in the success of personalized care delivery.
Even with AI adoption, there are interesting changes happening: investments concentrated in care delivery are now going to administrative simplification of operational and support functions, and this is something that we are very keen to enable.
What advice would you give someone starting their career in health AI?
Healthcare AI is an exciting industry slated to grow at almost 50% CAGR over the next 5 years. We are seeing an influx of bright minds come in to solve the most complex problems that the industry is challenged with. Not only healthtech companies, but also industry incumbents are using AI in innovative ways to solve for existing problems.
My advice to aspiring AI professionals in the healthcare industry would be that they should first seek to understand the industry really well. This will help guide their thinking and the context in which they operate while designing and developing AI solutions.