The Invisible Engine: How Operational Healthcare AI Is Redefining Hospital Efficiency

Artificial intelligence in healthcare is often framed through a clinical lens, focused on diagnostics, predictive models, and decision support.

Yet just beyond the patient room, a different reality begins to take shape.

Hospitals are not only centers of care. They are complex, constantly moving systems. Equipment flows across departments, staff coordinate under pressure, and critical decisions are made with limited visibility. Beneath all of this lies an operational layer that has remained largely inefficient for decades.

This is where a quieter form of AI is beginning to emerge. It does not replace clinicians; it enables them. It does not sit at the point of care; it strengthens everything around.

Bill Haughton, Vice President of Healthcare at Cognosos, speaks with CTO Magazine about what it truly means to bring intelligence into hospital operations, and why solving these invisible inefficiencies may be one of the most important shifts healthcare will undergo.

Healthcare AI and rethinking real-time visibility

At the center of this shift is a broader rethinking of how visibility works inside hospitals. Traditional approaches to tracking have often relied on heavy infrastructure and fragmented systems, making it difficult to achieve consistent, real-time insight across environments.

Newer models are beginning to take a different approach. At companies like Cognosos, this means combining widely available technologies such as GPS and Bluetooth with artificial intelligence to create a more continuous and accurate picture of how equipment and resources move through a hospital.

The value of this shift lies less in the technology itself and more in what it makes possible. When visibility improves, so does coordination. Equipment can be located more reliably, delays can be reduced, and staff spend less time navigating uncertainty.

Over time, this kind of operational clarity can change how decisions are made. Instead of reacting to shortages or disruptions, hospitals can begin to anticipate needs, balance resources more effectively, and reduce inefficiencies that have long been accepted as unavoidable.

To start with, Bill, you have spent over two decades in this space. How does your current role at Cognosos build on that journey?

Bill Haughton: I’m the VP of Healthcare at Cognosos. I have been in the tracking space for 20 years, tracking mobile medical equipment and pharmaceuticals. It is really a blend of tracking solutions.

I was SVP at Ekahau at the very start of the RTLS market introduction. So I have been in this space for that entire period of time.

When we talk about AI in hospitals today, the focus tends to remain on clinical intelligence. From your experience, how should leaders think about AI beyond that layer?

Mr. Haughton: We need to split AI into two conversations.

A lot of the time, when AI is discussed in a healthcare setting, it is around clinical decision support, much more on the clinical side. There is not a lot of conversation about how AI can help the operational side of a hospital.

That is where we are, behind the curtains.

We are part of the team that supports the care providers. Whether it is procedural areas like the OR or the med-surg floor, there is a whole team that supports the clinicians who provide care.

AI can come in, combined with tracking technology and what we do, to augment that environment and build a much more efficient system that directly impacts the quality of patient care.

If we stay on that operational lens, where do you see the biggest inefficiencies today?

Mr. Haughton: One of the biggest challenges hospitals have faced, and still continue to face, is how equipment is distributed and moved around.

The single biggest goal should be to ensure that nurses and care providers are never waiting for a piece of clean mobile equipment when and where it is needed. This has been a major challenge.

What we have seen over time is that hospitals tried to solve this by buying more equipment and staging it in every room, whether the room had patients or not. The outcome of not doing this well is that hospitals have massively overspent on equipment they do not really need, and it still has not solved the problem.

Hospitals are still at risk because this issue is not managed properly. More often than not, people talk about the symptoms and not the problem.

If nurses run out of feeding pumps, the assumption is that there are not enough. That is not necessarily true. It often means the equipment was not distributed or managed properly.

If there is an OR delay because infusion pumps were not available, again, it does not mean there are not enough pumps. It likely means the flow of equipment was not managed effectively.

You also see behaviors like hoarding. Nurses keep equipment nearby because they do not trust that it will be available when needed. That is a lack of confidence in the system.

That creates risks. The equipment might not have been cleaned properly. It might not have been serviced or could have missed a recall.

When you look at the core problem, it is very difficult to distribute clean equipment efficiently across a hospital, especially without the right tools.

Manual processes do not work. Staff are rounding, checking rooms, or hoping someone calls when supplies are low. That rarely happens because people are busy.

So the system breaks down, and hospitals end up in a cycle of overbuying and underutilizing.

That naturally leads to the question of differentiation. Hospitals have had tracking systems for years. Where does your AI-driven RTLS approach stand apart?

Mr. Haughton: One of the biggest challenges with tracking systems is accuracy at an effective price point.

If you are not confident about where something is, you cannot automate any process. If you are only 50 percent confident, then you are effectively zero percent automated.

Automation requires certainty.

The market over the last ten years has struggled with this. Highly accurate systems have been very expensive. Lower-cost systems do not deliver the level of accuracy needed.

What Cognosos has done is focus on delivering very high confidence across the entire facility, not just within a room but also in hallways, alcoves, and shared spaces, and doing that at a price point that makes sense.

We use AI as the core of our platform to deliver what we call ground truth. There are two ways to apply AI. One is to use it to improve weak data. The other is to use it to establish strong, reliable data from the start. We focus on getting the ground truth right. Once that is in place, you can automate processes with confidence.

Security and trust are critical in hospital environments. How do you approach cybersecurity and compliance?

Mr. Haughton: We take security very seriously, but we do not operate in the same way as systems that handle sensitive clinical data.

We do not access, transmit, or store PHI. Whether it is equipment or people, we are not transmitting that type of data. The RTLS industry typically associates a device ID with an asset or person in a secure cloud environment. For example, it could be hosted on Amazon Web Services.

If someone intercepted the data in transit, they would only see binary data. There is no meaningful information. We encrypt everything, and we work through cybersecurity requirements with hospitals, but we are very much on the edge of systems that handle sensitive data.

You mentioned earlier that hospitals often make purchasing decisions without strong data. How does your platform change that?

Mr. Haughton: Hospitals spend tens of millions of dollars each year on equipment. Many of those decisions are made without solid data. They are based on anecdotal input.

Nurses report shortages, which escalates through leadership, and the decision is made to buy more equipment.

What we do is provide hard data. We show what equipment is available, what is being used, and what is sitting idle.

The first thing we typically determine is how much excess equipment a hospital has. In many cases, it is between 25 and 40 percent. This is equipment that is not clinically needed. Without visibility, there is no way to know that.

Once hospitals have that data, they can make better decisions. They might redeploy equipment, share it across facilities, or avoid unnecessary purchases. Instead of buying based on assumptions, they can act based on evidence.

If a CTO is evaluating RTLS across multiple facilities, what are the key architectural and strategic considerations?

Mr. Haughton: There are a few areas to consider.

First, while there is a financial return, this is not purely a financial decision. It is about improving operational efficiency and clinical outcomes. Yes, you can eliminate unnecessary purchases and generate savings, but the primary goal is to support care delivery.

From a practical standpoint, you need full coverage of the facility and a clear understanding of how equipment moves through it. You also need the right teams to implement and operate the system effectively.

Ultimately, the focus should be on serving the frontline staff. If you can eliminate friction in their day and ensure they always have what they need, you are improving both efficiency and patient care.

The financial benefits follow from that.

Let’s make this tangible. If a hospital implements this tomorrow, what changes first, and what evolves over time?

Mr. Haughton: Within 30 days of going live, we are distributing equipment efficiently and eliminating equipment outages.

That is the first goal, ensuring 100 percent availability of clean equipment where it is needed.

Over time, we refine that and improve efficiency further. Within three months, we begin to show trends in equipment usage and potential excess. However, we would not make definitive conclusions at that stage because we need to account for seasonality. By 12 months, we have a complete data model that provides clear, evidence-based insights into what equipment is truly needed.

This eliminates guesswork in purchasing decisions.

There are also broader impacts. Improved infection prevention, better risk mitigation, and increased staff confidence. Nurses know they can rely on the system. From an operational standpoint, there are improvements in areas like central sterile processing and equipment servicing.

Another important area is staff safety. There is a growing concern around workplace safety in healthcare. If a staff member presses a panic button, the response must be precise and immediate. If responders go to the wrong location, even by one floor, that is a failure.

Accurate location data is critical for delivering real safety, not just the illusion of safety. Looking ahead, the future lies in connecting everything. Staff, patients, rooms, and equipment.

With that level of visibility, hospitals can enable smarter environments, automate workflows, improve capacity planning, and capture real-time operational data. These are goals the industry has had for decades, but they are now becoming achievable because of AI and reliable data.

For more perspectives on AI, enterprise systems, and digital transformation, explore here.