The $100M AI Question

Transcript

Imagine your organization approves $100 million for artificial intelligence. Not someday. Today. Every department has ideas. Operations wants automation. Technology wants new platforms. Finance wants savings. Customer experience wants virtual agents. And somewhere, someone has probably built a spreadsheet with 75 AI use cases ranked by enthusiasm. So where do you invest first? Actually, there may be a more important question: Where shouldn’t you invest at all?

Welcome to Healthcare Power Ups, the podcast that makes the complex business of healthcare a little easier to understand and a lot easier to talk about. Today’s topic is transformation. More specifically: What separates an AI experiment from an operation that has actually been transformed?

To explore that question, Healthcare Power Ups draws on insights from Madan Moudgal, Sagility’s Chief Digital Transformation Officer, and his perspective on what it takes to turn technology investment into meaningful business outcomes. And surprisingly, it doesn’t start with AI.

Start With the Problem, Not the AI. Go back to that $100 million. The obvious approach is to divide the money, fund a collection of promising use cases, launch some pilots, and see what works. But Madan’s starting point is almost the opposite. Don’t start with the AI. Start with the foundation. Because a $100 million technology investment does not automatically create a $100 million business outcome. Organizations need governance. Clear decision rights. Criteria for choosing which use cases deserve investment. People who can adopt new ways of working. And metrics that define what success actually looks like. But there’s another requirement that’s particularly important in healthcare: Domain expertise. Because when it comes to AI, possible and valuable are two very different things.

Consider a health plan contact center. A member calls to check the status of a claim. Could a sophisticated generative AI voice agent answer that question? Absolutely. But should it? Maybe not. A traditional IVR may already handle that simple transaction extremely well and at a lower cost. In that case, applying generative AI could actually make the economics worse. That’s one of the first traps of transformation: Using the newest technology instead of the right technology.

Now go to the opposite extreme. Give AI one of the contact center’s most complicated interactions. The member has questions involving benefits, claims, clinical information, and perhaps a provider issue. The conversation goes in several directions. The AI struggles. The member becomes frustrated. Eventually, the call gets transferred to a person anyway. Now the organization has paid for both the AI interaction and the human interaction and potentially delivered a worse member experience. The lesson isn’t that conversational AI doesn’t work. The lesson is that use case selection matters. Somewhere between a transaction simple enough for yesterday’s technology and an interaction complex enough to still require significant human judgment is an enormous landscape of opportunity. Finding it requires more than understanding AI. It requires understanding healthcare.

Automation Is Not Transformation. Here’s where the distinction gets interesting. Take any healthcare process and break it into individual steps. Retrieve a document. Move information from one system to another. Update a field. Route a request. There may be plenty of opportunities to automate those steps. And that can absolutely create value. But transformation asks a different question. Instead of asking: How can we make this step faster? Ask: Why does this step exist at all? What outcome are we trying to create? What happens upstream? What happens downstream? Is this handoff still necessary? And perhaps the most important question: If we were designing this process today, would we design it this way?

That’s the difference. Automation improves a step. Transformation rethinks the operating model. AI makes that distinction even more important because the number of things technology can do has expanded dramatically. But more capability doesn’t eliminate the need for judgment. It increases it. Sometimes the right answer is RPA. Sometimes it’s analytics. Sometimes it’s an AI agent. Sometimes it’s technology the organization already owns. And sometimes the right answer is still a person. The goal isn’t maximum AI. The goal is the best operating outcome.

How Do You Know It Worked? There’s another common problem with transformation initiatives. They can’t prove anything actually changed. A pilot launches. The demo looks impressive. People get excited. Someone calls it transformational. Then six months later, someone asks: What did it actually improve? That question is much harder to answer without a baseline. Before changing a process, organizations need to understand how it performs today. How long does it take? What does it cost? What is the quality? Where does rework happen? What is the member or provider experience? What is the financial impact? Then define the problem. And define success.

Maybe success means faster turnaround time. Maybe it means higher quality. Lower cost. Better experience. More revenue. Or some combination of them. The specific metric will vary. The important part is agreeing on the outcome before the transformation begins. Because a technology that works but never gets adopted is not transformation. A pilot that saves three minutes but doesn’t materially change the outcome is not transformation. And a brilliant platform nobody uses definitely isn’t transformation.

People, Process and Technology Have to Move Together. That leads to another important principle: People, process and technology have to change together. Healthcare organizations have said that for years. AI is making it unavoidable. Historically, operations and technology could function almost like separate worlds. Operations ran the business. Technology supported it. The technology team built something and handed it to operations. Operations decided whether it worked. But that model becomes increasingly difficult when technology itself is embedded in the operation.

The people who understand the healthcare workflow need to design alongside the people building the technology. And the people building the technology need to understand the business outcome they’re trying to create. That’s why Sagility’s approach increasingly brings together operations, technology, practice experts, transformation teams, and commercial teams. Not technology building for operations. Technology building with operations. And generative AI is making that boundary even less distinct. People who aren’t software engineers can increasingly prototype workflows, interact with complex systems, and use technology through natural language. The distinction between a technology person and a business person is beginning to blur. And that may be a preview of where healthcare operations are heading.

Meet Clients Where They Are. There’s another important reality. Large health plans already have technology. They have cloud environments. Data platforms. AI strategies. Vendor ecosystems. Governance standards. Security requirements. They aren’t waiting for someone to arrive with a black box and tell them to replace everything. Transformation partners have to meet clients where they are. That means being able to work across different technology environments and partner ecosystems. But technology knowledge alone isn’t enough.

Imagine a health plan has already selected its AI or cloud platform. The hardest question probably isn’t: Can someone install it? It’s: How should we use it inside this healthcare operation? Which use cases matter? How does the data work? Where does the workflow break? What can safely be automated? Where does human judgment still belong? And how does all of that fit within the client’s governance requirements? This is where healthcare domain expertise becomes especially valuable. Madan describes the opportunity as developing teams of healthcare technologists, people who understand both sides. Claims and cloud architecture. Healthcare workflows and AI. Clinical operations and integration. Technology and the business outcome it is supposed to create. That is a different kind of transformation capability.

Sell the Outcome, Not the Technology. And this leads directly to Sagility’s go to market story. Sagility is not trying to become a technology vendor. Sagility is a services company. But those services are increasingly technology enabled. That distinction matters. A client doesn’t usually arrive with a healthcare operations problem and ask for a menu of people, process, AI, analytics, and automation. They want the problem solved. So the answer shouldn’t be: Here’s the operations solution. Here’s the technology solution. Pick what you want.

The better answer is: Here’s the solution. Inside that solution might be healthcare expertise. Process redesign. Automation. AI. Analytics. Partner technology. The client’s existing technology. And human judgment. But the client shouldn’t have to assemble all those pieces. They’re buying the outcome. That’s why the technology story and the operations story increasingly need to become one story. If a client wants to understand the architecture, explain it. If they want to understand security, explain it. If they want to know how the AI works, show them. But the technology itself shouldn’t be the reason to buy. The outcome should be the reason.

So What Happens to the People? There’s one more part of the transformation story that can’t be ignored. The workforce. Over the next several years, AI will likely make more decisions inside healthcare operations. Not every decision. Clinical, regulatory, financial, and other high consequence decisions will continue to require appropriate human oversight and judgment. But many of the activities surrounding those decisions can increasingly be analyzed, prioritized, routed, recommended, or executed by technology.

That changes the role of people. Less time may be spent on repetitive work. More value may come from judgment, oversight, exception handling, and decision making. People will need to understand how to evaluate AI outputs. When to intervene. How to govern the technology. And how to add the context the machine doesn’t have. That makes transformation a workforce challenge as much as a technology challenge. Because the future of AI won’t depend only on whether the technology works. It will depend on whether people trust it enough to use it.

The $100 Million Answer. So, back to the beginning. Your organization has $100 million to invest in AI. Where should it go? There isn’t one answer. And that’s the point. The first question shouldn’t be: Where can we deploy AI? It should be: How should this operation work now that AI exists? Then choose the right combination of people, process, technology, and healthcare expertise to make that future possible. Because the goal isn’t AI. The goal isn’t automation. The goal isn’t even technology. The goal is better healthcare operations. Better experiences. Better economics. Better decisions. Better outcomes. Technology may be one of the most powerful tools healthcare has ever had to get there. But only when we choose the right problem first.

And that is today’s Healthcare Power Up. One piece of healthcare operations. Pulled apart. Put back together. And hopefully, a little easier to understand than it was before. Until next time, keep asking the question that tends to make healthcare a lot more interesting: What actually happens next?