Healthcare organizations have experimented aggressively with AI for several years. But in 2026, the conversation is shifting

The industry is moving beyond the phase of isolated demos and pilot projects and into something more difficult: operationalizing AI inside real healthcare workflows.

At Asembia’s AXS26 Summit, Infinitus CEO and Co-founder Ankit Jain sat down with Optum Rx CIO Santiago Abraham to discuss what healthcare organizations are learning as AI moves into production – from governance and quality measurement to organizational adoption and where AI is already delivering operational value.

Below is the first in a series of edited excerpts from their conversation. You can read Part 2 here, and Part 3 here. You can watch the full conversation here.

Ankit Jain: You’re sitting inside one of the biggest pharmacy benefit operations in this country, evaluating AI at a scale that most of us aren’t. What has the journey been like?

Santiago Abraham: Just in the past three months, we’ve seen an incredible amount of change in the maturity of the technology. So that really energizes us because it really highlights the opportunity ahead and really where we can go next with this technology.

I’m not a hype guy. I don’t do the hype cycle. I’m not going after the shiny object. But I will tell you that this stuff is looking really promising.

Ankit Jain: One of the things I’m hearing from across the ecosystem is that there’s a lot of board-level pressure right now around AI – pressure to increase revenue, reduce cost, and move quickly. 

How are you thinking about measuring risk when evaluating AI platforms and use cases?

Santiago Abraham: Governance matters in the healthcare space probably more so than in many other industries.

My comment there is, if the shopper agent makes the mistake of recommending the wrong thing to you, you added the wrong thing to the cart. If we make a mistake in healthcare with AI agents, there are big stakes at hand. So really quality is extraordinarily important in this space.

This is an area where there’s a ton of creativity and an abundance of opportunities. We have hundreds of use cases already in motion within Optum, so we’re bullish there. But ultimately the really critical step is: is governance appropriately in place? Do you have the guardrails in place? Do you have responsible use? Do you understand what that looks like?

And then secondly, from a quality perspective, are you managing the quality?

Ankit Jain: One of the things I always suggest people do when they test AI is test it for the same thing five times and see whether it behaves consistently five times. Most AI systems still won’t pass that very simple test.

Santiago Abraham: That’s the new term for essentially how we evaluate the performance of the LLMs and the applications that surround them.

You can build it, it looks great in the build, looks great when you’re testing it, but you deploy it into production and all of a sudden over time it’s not performing as well or it may be hallucinating. There are things in the environment that are changing, inclusive of the data sets, and getting that right is extraordinarily important.

Ankit Jain: Where are you actually seeing AI deliver value today?

Santiago Abraham: Let’s hit a couple: On the payor side of the business, prior auth is ultimately a point of friction. Really driving to improve turnaround times there is key.

So really supporting the clinician – the clinician is making the care decisions – but supporting that clinician to really be able to move through that in a more expedited, consistent fashion is super important to get patients to therapy faster. We’ve seen a 20% improvement just in the period that we’ve been running that, and we’re excited about the opportunities ahead there.

Let’s go to pharmacy. Pharmacy is an area where we also want to lean in with the technology. One of the first use cases we went after is quality-related events. Safety-related events are a big deal. We’re having AI apply consistency there and really address quality issues. Super important, really promising.

And then I’ll go to experience from a consumer perspective. I don’t know about you, but I am the first one that yells ‘operator’ at anything that comes at me from a bot and self-service perspective. I have zero patience for those things. But the newest generation of this stuff is materially better.

Ankit Jain: On the flip side, where has hype outrun reality?

Santiago Abraham: There’s a lot of full automation conversations happening right now – fully automated, especially patient- and member-facing experiences. In that space, we’re proceeding with caution.

Again, the technology is evolving pretty rapidly, so that is changing over time. But I think at the moment, if a vendor or partner is coming to you saying, ‘Hey, I’ve got a fully autonomous solution and I want to deploy it in your environment,’ definitely have some skepticism in that space and make sure you have the right quality guardrails established.

Ankit Jain: As you think about the next few years, where do you think AI will deliver the most value in healthcare?

Santiago Abraham: No. 1 is quality. Getting really opportunistic consistency with AI is really a game changer.

The second thing is cost. Everybody always goes to admin efficiency, but I’d go further than that: Really being able to get to insights at scale – fast insights at scale – and understanding where we’re going and what opportunities we have to really impact cost in a more productive way.

And then the third and final thing, and the thing I’m most excited about, is experience.

The experience of a hyper-personalized assistant that really is supporting you in your care journey, kind of walking you through that whole continuum, really taking care of you.

An assistant for the patient, but also an assistant for the pharmacist, an assistant for the operator, an assistant for our clients and customers.

This conversation has been lightly edited for length and clarity.