Episode 119: How AI Agents Are Giving Clinicians Time Back - Zyter TruCare CEO

For example, triage documentation and prior authorization are very written processes with a lot of business processes that can all be reimagined with agents very quickly and it's happening now, delivering some of that, but it extends this possibility without having to do a lot of modernization.

Now more than ever, you don't need to invest tons of money to modernize your systems to get value; you can start embedding AI in a thoughtful way and start to scale the process end to end.

Hi, my name is Demetri Panici and I'm a content creator, agency owner, and AI enthusiast. You're listening to the AI Agents Podcast brought to you by Jotform and featuring our very own CEO and founder, Aytekin Tank. This is the show where artificial intelligence meets innovation, productivity, and the tools shaping the future of work. Enjoy the show.

Hello and welcome back to another episode of the AI Agents podcast. In this episode, we are interviewing the CEO of Zider True Care, Sundar. How are you doing today?

Great to be here. Thanks for having me. Doing really well. Awesome. Well, I appreciate you for making the time on the show. Would love to hear a little bit just to kind of get started on how you got into the world of AI.

Sure. Way back when, I started because I was really good in math and luck would have it I got into a group doing computational modeling at MIT in physics and material science.

That's how I got into really thinking about how do you have a hypothesis, how do you model that, and how do you simulate real-world results either through models or digital twins.

You either prove your theory right or wrong and it led me to a deep belief in explainability. You could luck into many results but unless you had a clear basis to prove your theory, it wasn't useful.

My earliest exposure in AI was really exposure to explainability. If the simulation couldn't explain its reasoning, it wasn't useful. Since then, I've had many business experiences.

I got into consulting and did a lot of work in healthcare building behavioral and epidemiological models to simulate COVID progressions using digital twins of the US population.

The repeated lesson was that adoption was an explainability problem, not a data problem. That lesson stayed with me.

I've done a lot of consulting for Fortune 500s in digital transformation. The crux is how do you draw business value from any technology, not just AI.

There's a lot of massive tech spend now with AI, but in the last 20 years, there's been a lot of tech spend on ERP, CRM, RPA.

Consumer B2C technology evaluation has really helped with convenience, but on the business side, it's quite different. There's not as much value arriving on business processes.

That led me to believe that how you solve a business process is dead and how you solve broken workflows matters. Embedding AI doesn't fix broken workflows; it magnifies them.

That led me to healthcare and Zider, which serves clinicians as they hope to achieve their mission in serving patients and consumers.

We are an AI company with a mission to drive outcomes, not just features, functions, or point solutions.

I think that's a fair comment about workflows. If they're bad, they cause problems and they don't just get fixed with AI. Can you dive into that a little more, especially with what you guys are doing?

Absolutely. AI is going to be the technology of our life and a profound shift, but if you look at history, lessons repeat themselves.

Take electricity for instance. For three decades after its introduction, productivity improvements didn't happen because early factories swapped steam engines for motors but didn't redo the process end to end.

When electricity arrived, they just put it into the same flow layout, so there were no gains in productivity. It took Toyota years later to reimagine workflows with distributed cells and human-machine collaboration.

Transformation didn't occur from adding power; it came from rearchitecting the workflow. We're in a similar moment now.

Everybody's thinking about how to add AI to workflows, but the wrong question is how to add AI. The right question is how to solve for outcomes given what AI can accomplish.

If people had asked that with electricity, productivity gains would have come sooner. The lesson repeats with personal computing and cloud, where trust and reimagination were key.

Most people say transformation is embedding AI in use cases, but they're not fundamentally asking how to redo the work, which is hard because processes have been built over decades.

Our solution, Zider, serves 45 million healthcare members with platforms for utilization management, care management, and population health management, with deep expertise in clinical workflows.

Our agentic AI, Symphony, creates modular orchestrated systems of AI agents that solve end-to-end workflows, not just one-off use cases.

Embedding AI in isolated tasks adds to tech debt and doesn't transform processes. We create multi-agent systems that orchestrate workflows and keep humans in the loop to ensure trust.

I really like that story. Thanks for sharing it. I'll keep that in mind regarding adoption.

AI exists, but many companies struggle to realize value. How did you personally find your way into the intersection of AI and healthcare, and what motivated you to tackle healthcare complexities?

Great question. Maslow's hierarchy of needs shows that fundamentally everyone wants a better healthcare experience to live longer and healthier lives.

Healthcare is siloed with data sharing issues, unstructured data, regulations, and privacy concerns. Modernizing old technologies requires significant investment many can't afford.

We're in an enormous affordability crisis with healthcare spending growing unsustainably, and consumers are unhappy with the system.

I've been in healthcare for 25 years and see a 90% value gap because technology spend doesn't rearchitect business processes.

The issue is human imagination and adoption, not technology. Past technologies solved small use cases but had little meaningful impact in a complex ecosystem.

If we look at end-to-end healthcare, the answer is how to get the best data to predict next steps and how AI agents help humans take timely, meaningful actions.

Agentic AI embeds cognition in every node, allowing autonomous problem-solving within defined boundaries, enabling rethinking of processes end to end.

For example, triage documentation and prior authorization are written processes with many business steps that can be reimagined with agents quickly, extending possibilities without heavy modernization.

There's still burden on data readiness and architecture for safety, but now you don't need to invest tons of money to modernize systems to get value; you can embed AI thoughtfully and scale processes end to end.

That's exciting for healthcare because there's tremendous value to unlock.

Absolutely. There's a definite need for AI to help people in healthcare because many struggle with costs and care.

Premiums may be manageable when young, but as people age and have families, costs become wild. Healthcare is a good place for AI to make universally good waves.

What are you practically doing to help in healthcare? Could you dive deeper into your product and what you do?

Utilization management or prior authorization is a common process where approvals require back and forth between insurers and providers, often involving fax and manual reviews.

Clinicians review eligibility, benefits, and medical necessity, often needing missing information from providers, causing delays and administrative burden.

AI can help by deploying 40 agents working on behalf of one clinician to streamline this process, improving efficiency and turnaround times.

AI works in tandem with humans with responsible AI layers; humans review summaries, eligibility, benefits, clinical criteria, fetch missing information, and make decisions.

This workflow is 80 to 90% more efficient, turnaround times improve multifold, and patients get what they need promptly.

What does this give clinicians time back for? Clinician capacity is a significant issue; more clinicians are needed to focus on keeping members healthy, not overwhelmed by volume.

Automating administrative work allows clinicians to engage more with care management programs, helping patients stay healthy with aligned incentives for insurers and providers.

For example, I use a digital therapist for back spasms, which monitors compliance and provides feedback, allowing the physical therapist to support many patients effectively.

Similar to self-service checkouts with humans in the loop, care teams include AI agents, nurses, and clinicians working together, with AI often in front communicating with members.

This model untaps clinician potential aligned with member goals and helps many patients receive support.

Clinicians often feel like they're playing catchup with administrative tasks. AI can relieve this burden, allowing them to focus on their core job.

How does the market handle compliance with AI? Trust is key; humans must be in the loop for clinicians to accept AI recommendations.

Studies show consumers don't reject AI but need explainability and a human in the loop to accept AI-driven systems.

Compliance requires high security and privacy. For clinicians to trust AI recommendations, three things matter: human in the loop, explainability, and confidence in recommendations.

Low confidence predictions get overridden 99% of the time, but with transparent explainability and high confidence, overrides drop to 1.7%.

Better systems show rationale, evidence base, and present it simply so clinicians don't have to research explanations themselves.

Showing multiple models converging on the same recommendation and repeated questioning increases confidence and compliance.

Where do you see AI agents heading in the next few years? AI agents are doing valuable things now, but the future holds significant potential.

We are at a cusp of a significant singularity or dislocation from an AI agent perspective, where friction and coordination costs in business processes will approach zero.

End-to-end processes can become 80 to 90% efficient, making decisions more instantaneously, and workflows will be reimagined very fast through agentic AI.

Even with modest assumptions about cognitive progress, reinforcement learning from human feedback will scale unprecedentedly, driving costs toward zero.

Businesses adopting this reimagination journey will create extraordinary value propositions with low prices or significant benefits for consumers.

Software engineering is a big component, with flagship models like Gemini 3 Pro and Opus advancing rapidly and reducing pricing significantly.

The concept of a co-pilot helping with coding is evolving into autonomous decision-making that can recursively solve problems and create products.

This productivity gain is extraordinary and must be managed responsibly with stress testing and oversight.

What is your company's vision for the next five years? We want to help businesses reimagine workflows and be the Toyotas of the AI revolution, not those who just pour AI into old processes.

We provide the Symphony platform and sandbox for businesses to learn, create agents, and build workflows to achieve the best outcomes continuously.

While we started in healthcare, we aim to help clients in regulated, complex industries with multiple stakeholders and friction to reimagine processes globally.

Advances in AI like generative AI and world models will accelerate this, and we help businesses bring lab advances to real-world processes responsibly.

What is your personal favorite AI tool? I like ChatGPT a lot. It understands me well and offers a high level of personalization.

I have a newsletter called Student of Life about AI leadership and parenting. ChatGPT helps with bedtime stories, parenting advice, and more, showing impressive personalization.

Where can people find out more about Zider? Visit zida.com to learn about our product Symphony, which integrates with existing digital tools to reimagine processes and agentify workflows.

Thank you for having me. I appreciate the discussion and your work. AI won't transform your business just by buying the right tool; it transforms by redesigning systems around it.

Thanks for listening. If you liked this episode, please like, subscribe, and leave a review on Apple Podcast and Spotify to support the show.