AI-Powered Financial Optimization in Healthcare

Venkat, could you provide some background?

I've spent the last 15 years thinking about how we innovate in healthcare. First, I was really thinking about how we move away from fee-for-service care and American healthcare to value. I helped start a company inside DaVita called Paladina, focused on direct primary care. I got to do global work. In the second chapter, I worked in Europe, the Middle East, and Canada for a research data consulting firm called the Advisory Board, where I got to see how not just innovation is done here in the US, but genuinely around the world. Ten years ago, I got really deep into applied machine learning work and AI work in healthcare. I got to see a company at the founding stage, Qventus, applying machine learning to help think about hospital operations. Right before Midstream, I spent time at Andreessen Horowitz, which is a large venture fund, thinking about the future of healthcare.

Midstream is a financial action platform. It's AI native. It’s as complicated as ever for a health system to manage, procure, and negotiate with multiple stakeholders in a world where there's a lot of data, a lot of meetings, and a lot of dashboards. We help create clarification, and most importantly, think about making things much more proactive than reactive. It's been amazing thinking about how we go from a reactive world to a proactive world? How do we go from a dashboard to an actionable next step? And it's been phenomenal partnering with an organization like Common Spirit to show the world that this is not only possible, but there are incredible outcomes with it.

Could you tell me how this platform works and the benefits it provides to healthcare organizations?

In today's world, there is so much information, so many sources where data comes from. There are unstructured datasets, like contracts, for example, where it might be sitting on someone's desk, in a PDF, in a big repository of files, or external data on the internet. There's structured data. Maybe it's sitting in an ERP. Maybe it's sitting in another system of record, like an EMR. Then there is context in people's heads. If you take all of that, analyze it, and present it to a meeting somewhere, people make decisions, and actions take place. Instead of that, what Midstream does is sort of sequences all of that data, unstructured, structured data, even publicly available data. It takes context from the people working behind the scenes in the administrative layer, does the analytics in the background, and then proactively says, "Hey, here's a next step."

It's sort of going from the era when people used to print maps and hoped for a good sense of direction to Google Maps, which is a lot more prescriptive. I think that paradigm hasn't been really possible in financial decisions.

What Midstream does, for example, when you buy a really expensive device, it looks at not just the contract, but it also looks at the financial data, and says: “Hey, in the moment, we've identified some waste that could occur. Let's go ahead and remedy it. Are you interested in this particular next step?”

In our industry, there are actually many ways to save money without changing workflows or people's jobs. There are just decisions that are sitting on the table that, if you had gone through the data, you would identify that it happens today in a very manual way.

Could you tell us how this platform may differ from other platforms?

Why we're so excited about this step function change is that the system actually sequences all of the data from the contract. It's based on market share data. It's based on volume; it's based on lots of things. It's able to connect the dots between the contract, which is a static PDF, and the transactional data and all the decisions that are being made.

There are some phenomenal outcomes where we've shown that by taking simple next steps that are already in the contract, you can save dollars for the industry. There's a long affordability crisis in healthcare. Healthcare is too expensive.

Can you elaborate about the shift from analytics to agentic action, or AI?

There is a lot of complexity in devices and drugs, and a lot of decisions are being made. It's very cumbersome. The agentic world says: “Hey, let the agent read all the data. Let the agent analyze the data, and instead of having 10 meetings, let me just ask the agent questions about the data.” Now, if you want to understand the why behind it, you can choose to do that on your own time. That's why agents are so interesting. It can go from just cruise control and lane assist to more of an augmented copilot. It's that sort of step-by-step of how do we take all these little decisions and let people focus on the big ones, where it shouldn't be a human's job to read 800 pages

Could you tell me a bit about the partnership with CommonSpirit?

CommonSpirit is one of the largest not-for-profits in the country, and its mission to advance human kindness aligns with ours. Our mission is about creating zero waste in American healthcare and around the world, and maximizing savings.

For us, this is not just a one-off pilot; it's a long-term partnership where we want this to really apply in many different areas. We started with supply chain, but we genuinely think that if we do this the right way, others will follow. Other health systems will follow. Because it's not just a one-off hospital, you know, it's to, you know, hundreds of hospitals. Their leadership team is incredibly transformational and forward-thinking, which we genuinely enjoy.

When did you start implementing the system?

We started implementing this in a big way last year. We really got to implement and see the results in 2025.

Could you speak to some of the results you have seen?

We're very proud of our outcomes. First, we think that in an ocean of AI companies, leading with results really matters because that's what people really care about at the end of the day. And so, we're very proud of it. We think a lot about speed-to-value.

Can you talk about some of the challenges you faced and how you solved those?

I think people think a lot about agents and AI, but it's the data story that feeds into it that makes a great application. That's step one. Step two is…it's about the people. How do people change their decisions? The last thing is that the pace of change today is so dramatic that every three days in AI feels like a new year.

Thinking about that, what do you foresee for the future?

For us, all of this is about how you achieve outcomes and how you reduce waste. I think there are a lot of people in our industry who are very burnt out. I don't think we talk about operator burnout and administrative burnout, because healthcare is so complicated in America. There are so many decisions to make, from little decisions on purchasing or execution of a contract, to thinking about how new drugs, devices, and technology are changing. I think that is probably what I'm most excited about: this intelligence layer that's proactive and assistive and can connect the dots in a very siloed environment, in a very big healthcare ocean.

How has your international experience helped you with this development?

It's a great perspective. I think innovation in Silicon Valley and for healthcare is about U.S. healthcare, because there are the complexities and the incentive models. They're so complex that you spend all your energy here. I think we have an affordability problem around the world. When we think about our mission to achieve zero waste in healthcare and maximize savings for society, I think this can really scale across the globe. We are really thinking about financial decisions, and I think this is universal.

What is your advice for healthcare leaders?

I grew up in India and Silicon Valley, and I genuinely think AI is a technical revolution. I think it's a platform shift. I think it's genuinely going to be great for mankind, harnessed the right way. But we have to think about the humans on the other side who are using these things and approach it with an understanding of all the other aspects of their workflows, their priorities. We have to be thoughtful of governance.