Process mining and process intelligence are emerging transformative technologies that offer organizations, particularly those in healthcare and service sectors, the ability to streamline processes, reduce costs, and enhance performance. Yet many leaders are still unaware of their full potential. McKinsey Talks Operations host Daphne Luchtenberg is joined by Mike Kaufmann, a healthcare leader, board member, and former CEO of a Fortune 15 company, and Steven Eklund, a McKinsey partner specializing in service operations. Together, they bring their extensive knowledge of optimizing processes across service industries to shine a light on technologies and how they add value for healthcare and service organizations.
The following conversation has been edited for length and clarity.
Daphne Luchtenberg: Welcome to you both.
Mike Kaufmann: Well, thanks for having me. It’s great to be here.
Steven Eklund: Good day, nice to be here and thank you for the opportunity to share some of what we’ve learned.
Daphne Luchtenberg: Mike, you spent many years in various roles at your last company and got to see many aspects of that organization. Once you stepped into the CEO role, I’m sure you had a good sense of what was working and what could be improved. Were you already aware of the power that process mining could bring? What prompted you to start looking into that discipline?
Mike Kaufmann: It’s a great question because I had never heard of process intelligence or process mining—at least in a way that I understood what it meant, and what the power of it was. I’d spent 20 years in the C-suite in different roles, with the last five as CEO. But one of the things I learned really early on as a C-suite leader is that you need to be very purposeful about setting aside time to have clean, clear, quiet time to think strategically and to network.
And I think a lot of times people are not purposeful about that, and at the end of their day, weeks, months, years, they find out they didn’t really get to do that. And so I was always very purposeful about that. And while I was taking advantage of a networking situation, I started talking to another well-respected CEO who was running a large company. And he started telling me about process mining and process intelligence and how they were saving his company over $100 million a year. To me, that was enough to be curious, ask more questions, meet with the company that he was utilizing, work on a proof of concept with them, and we found there was significant value for us at Cardinal Health. And so that’s how I became part of it, got deep into it, and I found it to be incredibly valuable for us.
Daphne Luchtenberg: And what kind of business challenges did you put under the microscope once you discovered this new way of bringing transparency?
Mike Kaufmann: One of the things I learned early on about process mining and process intelligence is you can think about it in a couple of different buckets. First of all, it’s got the power to give you immediate savings. You could run it in areas like accounts payable, and it’ll see where you paid something twice or paid it early or put a customer on credit hold or when something in the supply chain isn’t going to come in on time. There are a lot of opportunities to deploy it in areas that immediately affect your bottom line.
But what’s just as powerful, or maybe could be more powerful depending on your situation, is when you can use process intelligence effectively to cut across multiple systems within your company where you normally don’t get to see end to end an entire process and how it works. You’re able to gain insights into how your customer is buying, how your margins are being generated, and how you can improve customer satisfaction. You start to gather insights that you can act upon and do other things in order to improve your net promoter scores, your supply chain, your relationships with suppliers.
Subscribe to the McKinsey Talks Operations podcast
Daphne Luchtenberg: So that must have been massively eye-opening for you when you started to apply that to Cardinal, right?
Mike Kaufmann: It really is, because when you have a company that has multiple processes, maybe multiple acquisitions, or complicated connections between various systems, then there’s most likely going to be some problem in that process that never works A to B to C to D like you’d map it out and design it. It sometimes goes to G, and process intelligence and process mapping and mining really help you see how processes really work, so that you can either fix the problems that might be happening or you can connect them and get insights into your business that allow you to drive better results.
Daphne Luchtenberg: Thank you for explaining that so well. Steve, let me bring you in here. You have worked with numerous clients across the services sector. What are some of the business challenges that you’re seeing process mining being put to?
Steven Eklund: I’m going to piggyback a little bit on what Mike shared. We see three kinds of common challenges that most organizations face as they work through process outcomes and process mining and insights. The first is that they’ve historically thought about themselves in functional silos. “I’m focused on finance and accounts payable, not the procure-to-pay process.” Or, “I handle orders, make credit decisions, and collect cash from customers,” but not necessarily “What are the outcomes I’m driving across a process like order to cash?” And so for us, one of the challenges is breaking that functional silo and starting to align the organization, not necessarily in boxes and lines, to say you now are organizing yourself in a new way that’s end to end but really to drive clarity, transparency, and accountability across a set of business outcomes linked to those end-to-end processes.
Concepts like business process owner—someone who’s going to facilitate the insights that come from a tool like process mining and direct the organization to drive toward a better outcome, whether it be customer experience, effectiveness, speed, quality, accuracy, or productivity—is a critical role that now has to be in most organizations.
I think the second point is, and Mike touched a little bit on this, some of the technology barriers, right? Many organizations have a mix of multiple enterprise-resource-planning systems, manual interventions, and other tools and capabilities that they’ve invested in to work around blocks in their process flow and to try to streamline some of the labor of the process. And what we’ve seen is that it’s really hard for organizations to see through that mess, if you will, of technology, and understand exactly how a transaction flows and what all the blockers are.
The process mining tool allows us to—Mike, I’m going to steal a word from you—“curate” that process, organize it in a way, and really visually understand what’s happening and some of the changes that could free up manual work and improve the outcomes. So the technology barrier would be number two. And I think the third is the data.
Much of what we’re focused on here is how we not only curate the process but make sure we have transparency around the data and manage and keep the data clean over time. Oftentimes what we’ve seen is this uncovers not only process inefficiencies but opportunities for better data management and governance on how you organize that data for the organization, not for a discrete function to take advantage of it.
So those are three common challenges. I think these challenges are particularly relevant today. Everyone’s talking about AI, agentic, and gen AI. In order to take advantage of some of those newer technologies, you have to build the foundations around your process and your data, starting with process insights, to be able to organize and leverage those newer technologies.
Daphne Luchtenberg: I was just going to ask you another clarifying question. As you work with clients, and as you bring this kind of transparency and data mining into it, I imagine it has a bit of a democratizing effect, in the sense that it’s bringing a version of the truth that can sit across teams, so it’s no one team not doing their work or being challenged because the process isn’t working or it’s not optimized, it’s like the entire organization. Do you find that that’s a unifying factor, or are there other things that need to happen for teams to look at that objectively?
Steven Eklund: It is a unifying factor, but it requires a different approach than what most organizations have done: they’ve thought about transforming their front, middle, and back office functions and processes. And it starts with the CEO—Mike, as an example at Cardinal Health—and an executive leadership team that is unified across a set of business outcomes we’re trying to accomplish, and organizing the work, and how we’re going to drive change in those cross-functional teams to enable the outcome “unlock” that the executive team has set out to.
It starts with how you approach some of these changes, then putting in place resources and insights that allow us to ask not only “How do we fix what’s broken?” but “How do we manage and maintain the improvements that we have over time—and frankly, build the discipline in the organization that says where I am today needs to continuously be challenged and proved as we go forward?” Because the technology, the data, continues to evolve in a way that requires that level of change over time.
Daphne Luchtenberg: Mike, let me bring you back in. So what was your experience and what did you learn?
Mike Kaufmann: I want to emphasize something Steve said that I thought was really spot on because one of the things that I’ve learned early on in both my experience as a customer of process mining and process intelligence, and it’s even expanded since then, and working with these companies going forward and with customers that are utilizing it, is that for it to be successful and add the most value, you have got to look at the end-to-end process.
A lot of times, what happens is a person who only has a piece of the process gets designated as the lead on the project. And you can never get the full value because you get stuck in the area that they are responsible for, which is only a piece of the process. And so one of the things I’ve learned is that if you don’t have C-suite sponsorship—typically chief information officer, COO, CFO, or CEO, because I was a sponsor at Cardinal—if you don’t have someone at that level who cuts across all processes and periodically checks in—and when I say sponsor, I know it’s unrealistic to think the CEO is going to spend ten hours a month on it—but if you just check in once a month for an hour and say, "Tell us what progress we’ve made, that’s the type of energy that gets it going."
You really need to have someone senior that says, “this is important; we’re going to look at it end to end,” or you get trapped in silos. And I’ve seen dozens of companies that are dramatically underutilizing the potential of process intelligence and process mapping and all those things that come along with this because they have no senior sponsorship, and it gets stuck delivering two or three million of value for $200,000 of cost, which is a great ROI, but it’s in one little area, versus leveraging and making $200 million of value because it gets across the entire process in 25 different areas. So I do think senior sponsorship is a key piece of it.
Second, what I’ve seen to be successful is that it needs to be business led, IT supported, not IT led, business supported because if you don’t have the business lead, it won’t get the same level of attention and input and desire to achieve the final results. IT is critical to it. They’re important. They can bring a lot of value, but it should be business led. So those are two things I think that will really drive the success of someone in the areas of process mining or process intelligence.
Steven Eklund: Business led, IT enabled is critical to my first point, right? Which is, again, these are business outcomes we’re trying to drive. We’re trying to improve the straight-through processing, working capital terms, the productivity across the organization. I think that’s a tough thing to solve when it’s just a technology solution. That’s critically important as you think about what we are trying to achieve here with the foundation of process mining.
Daphne Luchtenberg: That comes back to Mike’s advice that you need to have senior leadership that keeps anchoring everything back into that business horizon, right? Mike, can you say more about how you, in your organization, overcame some of the some of the cultural challenges and people challenges? How did you get your teams to embrace the new way of working, and this new way of innovating?
Mike Kaufmann: It’s a great question because not only did I see it where I work, but I’ve seen it at many companies. Again, if you’re not setting goals and periodically challenging the team to find new use cases, it will stall. The other thing you have to remember that’s important is that, oftentimes, the IT department’s solution—and I get it, it’s where they came from—is to build it themselves versus buy it from somebody else. And a lot of times, if you turn it over and the process intelligence/process mining finds a problem, the solution can sometimes be, “Well, why don’t we just build it ourselves?” which could be ten times or more the cost of just utilizing the process mining or process intelligence, and can elongate the process for very long periods of time versus getting value right away.
So I think you need to make sure that as a leader, you’re making sure to tell everybody, “It’s OK to be transparent with our processes. Show me where the warts are, and let’s fix them as a team”—a real positive message. It’s important that you deliver that, so people want to embrace it, because it is going to find issues within your company. Second, you need to have that sponsorship to see it end to end, so it doesn’t get trapped.
Third is you need to set goals. Once I did the proof of concept, we set a goal the first year of, say, 30 million of value from it. And then, in year two, the goal was $70 million of value from it, because we saw the power of it. And just to go in and say, “Let’s do the best we can with it,” it just doesn’t drive behaviors. But when you hold people accountable and say, “This year we’re going to use this technology to drive $100 million of savings,” people wake up and they start thinking of use cases, they start making sure you cut across the entire company, and they really get after it in order to achieve the results. I think that’s why, again, I emphasize the importance of a C-suite leader who has the ability to set those goals and hold people accountable. When you do that, it unlocks a tremendous amount of value in the organization.
Daphne Luchtenberg: And I imagine a lot of time was spent also celebrating those successes, right? Those kind of early milestones and quick wins.
Mike Kaufmann: A hundred percent. And not only is it just celebrating the wins, but it’s discussing where you used it and it didn’t work. Because sometimes you use it, and you find out that maybe you are better in that area than you thought you were. Or maybe it delivers $2 million of value or $100,000 of value, and you really need to go after the stuff that’s five times that. So there are times that you celebrate the wins, but you also talk about when it came up short, why, and where you can go. It’s a really powerful tool, but it’s not always the best tool in every situation.
Daphne Luchtenberg: How would you now coach other peers and other leaders of businesses to get started? What are some of first things, Steve, let me come to you first with that one. When you’re talking to your clients, how do you advise them on how to get started first?
Steven Eklund: I would say it’s first recognizing this is not a novelty, right? I think with the examples that Mike shared about the impact and where we’ve seen it across a whole host of companies, we’d say the proof that it can demonstrate and realize value, starting with understanding the process, I think has already been clearly articulated, regardless of what process you’re talking about. For me, the key is really to get started. To pick a technology tool to work across all the requirements about data and security and to align the organization to pick a place to say, “Where do we want to challenge ourselves to get better?” and to define what “getting better” means.
Then use that as the lighthouse example for celebrating the early wins but also proving that this is a capability and a muscle that we need to build across the organization. We’re going to dedicate executive sponsorship. We’re going to align project leaders to it. We’re going to talk about the outcomes, the learnings, the good, bad, and indifferent, on a regular basis. And we’re going to quickly put in place interventions if we don’t see we’re achieving the outcome we want. I think the key for me is to get started in an area across some of the processes we’ve already discussed—order to cash, procure to pay, hire to retire, you know, as examples—that would be number one on my list.
Daphne Luchtenberg: And Mike, what about you? What gets the light bulb on in the people you talk to?
Mike Kaufmann: I believe that if you fully implement process mining or process intelligence, you can improve your bottom line 3 to 5 percent. If you’re $100 million bottom-line company, there’s probably at least five million involved in that with a ten times ROI. And if you’re a $2 billion company, I believe there’s at least $100 million there, and maybe more as you get complicated—and also maybe more if you include the insights and things, right?
I think what happens is C-suite executives often spend their time on things they believe add value to the bottom line. So they prioritize their time based on what adds, right? Makes sense. That’s what I did. I think what helped me was when this other executive that I admired and respected told me it saved $100 million to their bottom line. It gave me the scale in my mind that I should prioritize this. And I think a lot of reasons why it can’t get on the radar of the C-suite is they don’t have enough confidence that it’s going to deliver a 3 to 5 percent, because what CEO or CFO is not going to want to prove their bottom line by that much without doing anything like cutting people?
The answer is probably very few wouldn’t say they would want to spend time on it. But very few do spend time on process intelligence and process mining because they don’t understand it at that level, and they don’t have the confidence like I did that says, “Wow, this is a really well-run company. He just told me they saved $100 million. This is something I’ve got to prioritize.” I think as more and more people learn about it and build confidence and realize the scale at which you can drive value to the bottom line, you’ll see more and more sponsorship, more and more positive results.
Daphne Luchtenberg: No one ever stands still. And obviously, technology is evolving. What is next on the horizon for process mining? Steve, you first.
Steven Eklund: I think we’re at an inflection point where, by re-orchestrating how work happens across an organization, whether with AI, gen AI, or agentic, that’s going to come in the next several years, and in many cases are here today, we think we’re going to see tremendous productivity and better outcomes for core processes across every organization. We think the numbers are going to be quite staggering in terms of the productivity unlock across those organizations and those processes.
And so for us, it’s how we leverage the process mining and process insights today to understand how we work across multiple systems, data repositories, and organizations and start to understand what the outcomes are, how good we are, or how healthy we are in those core processes and leveraging the technology to think bigger. It’s Mike’s $100 million goal but sprinkled across a whole host of different processes in a way that is going to allow the organization to work faster and be, in many areas, quite leaner.
And so I think the process mining of today is really about building the foundation of that aspiration and is going to allow the organizations to work through all of the reasons why they couldn’t do it with automation and other technologies that have been around for the last five years. We would say this is a necessary prerequisite to get to that aspiration and that level of productivity and effectiveness.
Daphne Luchtenberg: And Mike, with your hands-on experience, where do you see this going? Where is the potential?
Mike Kaufmann: I think it’s well established in the areas of driving direct savings and providing insights. Where people are yet to really utilize it is where Steve’s going, around the area of AI and how you can effectively use AI.
Every CEO’s C-suite executives are getting pressure. “Where are we doing AI?” the board is asking everyone. What are you doing in this space? Every single board meeting. The problem is, right now, you can generally effectively use it on contracts. You can run a contract through a system, and it can tell you what areas you need to improve. And there are certain areas where it’s used well.
But where it’s really tricky is that when you go to apply AI to your actual financial data or supply chain data, when that data is not connected, curated, and cleaned in a way that AI can tap into it and actually give you the right answer, it’s dangerous to use. And so I think why people are really struggling to use it outside of some of those areas is because the data’s not organized or connected in a way that you don’t have a lot of risk of getting a hallucination or an incorrect answer from the data. And so without process intelligence, process mining, connecting all of it so that you can see data across multiple systems to give you the right answer, there’s really no way that you can effectively do AI in your company, because most of us know our data is not in perfect, mint condition to give you answers.
The second thing is, even after you do that, and let’s say for a minute, somehow you do use it effectively. You apply it. But you cannot have AI unchecked. You need to monitor that the AI result is accurate. And again, that’s where process intelligence and mining are going to come into play, where it monitors the actual results you’re getting to make sure you don’t do something that’s so incorrect, or damages the customer relationship, or hurts you financially.
So I think the importance of process intelligence in the area of AI is dramatically underestimated, and I don’t think companies have yet fully embraced the understanding of how they’re going to have to step into process intelligence before they go there. Because in my mind, there’s just no AI without process intelligence.
Daphne Luchtenberg: That’s a great way of closing this conversation. And Mike and Steve, we’re going to have to get you back because this is evolving quickly, and our expertise—and also how we’re seeing business cases and the case studies themselves evolve in different organizations—will grow as well. We’re all still very much learning in this space. Thank you both so much for joining us.
