Discover how AI is reshaping asset management, from smarter workflows to data-driven efficiency, and why clean data is key to unlocking real ROI.
In this episode, IBM’s Kal Gyimesi and Vishal Rane explore how AI is redefining efficiency in asset lifecycle management. They share why strong data governance is the foundation for successful AI adoption — noting that only 25% of AI projects meet ROI expectations due to disconnected systems. The conversation breaks down different types of AI, including generative AI, machine learning, chat assistance and intelligent agents, and how each drives smarter decision-making. They also emphasize the need for continuous testing, monitoring and responsible deployment to ensure reliability and long-term value across operations.
00:00 Introduction to AI in Facility Management
01:38 The Hype and Reality of AI
03:10 Types of AI in Organizations
04:34 Asset Lifecycle Management
07:18 Data Governance and AI Implementation
08:31 Corporate Real Estate and Digital Transformation
12:09 Conclusion and Key Takeaways
This episode is sponsored by ODP Business Solutions!
Kal Gyimesi: [00:00:00] What we wanna do is add AI as an addition into workflows that we can make more efficient. What we don't want to do is take a mess of systems and think that we can solve problems by throwing AI on top of it.
Host: Welcome to Connected fm, a podcast connecting you to the latest insights, tools, and resources to help you succeed in facility management. This podcast is brought to you by ifma, the leading professional association for facility managers. If you're ready to grow your network and advance in your career, go to ifma.org to get started.
In today's episode, IBM's Kal Gyimesi and Vishal Rane explore how AI is redefining efficiency and asset lifecycle management. They share why strong data governance is the foundation for successful AI adoption. They also emphasize the need for continuous testing, monitoring, and responsible deployment to ensure reliability and long-term value.
Now let's get into [00:01:00] it.
Kal Gyimesi: My name is Kal Gyimesi. I'm on the product management team at IBM for our asset lifecycle management solution my partner in crime here.
Vishal Rane: Yeah, I'm also on the product management team for our asset lifecycle management solutions. My name is Vishal Rane. And little bit of background about me. I came from industry. I have a IFMA accredited degree in facilities management. And I transitioned from public sector to IBM about a couple years ago.
And you're an FMP? I am an FMP. That is right.
Kal Gyimesi: There we go. All right, so. You know, a AI has been a hot topic and I see it all over the place here, and I've already seen it mentioned in a number of uh, presentations and I see it in a whole bunch of booths. There's endless amounts of hype and there's endless amounts of caution as well.
So, you know, we see eight outta 10 CEOs. We talk to a lot. We do a lot of big studies around all the C-Suite all you know, [00:02:00] CEOs, CFOs, CIOs. Chief Operating Officers eight outta 10 are pushing AI fuel cost savings, and we have it throughout IBM as well. We have a challenge in our internal operations to save by the end of 2025, be able to save $4.5 billion through AI enabled processes and AI enabled efficiencies.
61% of CEOs think that competitive advantage depends on who has the most advanced. Generative ai. But at the same time, , we've also seen some studies come out that shows the reality check that a lot of the projects that AI takes on maybe isn't given the ROI isn't giving the payback that people are expecting.
So you have to be a little bit careful with it. It can't just go into, just apply to any type of, system situation that you have, you often have to clean up your systems first and do the blocking and tackling and the foundational work before you apply AI to some processes. Only 25% of AI initiatives [00:03:00] have delivered the expected ROI and 50% of CEOs say that disconnected and poor systems are the root cause of all that.
So when we think of what types of AI go into, in into big organizations. We break it down into four areas. One is generative ai. That's AI that's able to generate new content. You guys are probably using it when you are using chat, GPT or Gemini or copilot in your own personal lives, and maybe you're using it at work as well.
Machine learning is the second one that's AI that predicts, that uses historical data and statistical and in-depth statistical analysis to fill in. Some gaps and some interpolate and extrapolate to use machine learning and deep learning to solve big problems.
Second third is ai, the chat. So AI assistance. We probably come into contact with those and those can be really hit or miss if it's. Really trained on good company data it could be a [00:04:00] effective assistant, and that's what we've started to build into our suites. But we've also seen customer service bots that don't, aren't able to answer anything but the most rudimentary of questions.
And they get you quickly frustrated. And if the organization's put in processes where they've cut you off from a human interaction, if the AI isn't working, then that becomes a source of frustration and then. The next thing that's coming down the pike is AI agents and a, the agents are things that will begin to do the work for you.
We call it agentic ai.
Vishal Rane: Jumping a little further on that, traditionally when people think about asset management, they're thinking about operations and maintenance and Maximo, but we're broadening that perspective when we talk about asset lifecycle management.
Right. We're talking about capital planning, project management, obviously operations and maintenance, energy management, sustainability, lease management for those assets that require that, that aspect of asset Lexi [00:05:00] management.
You know, everyone in this room with how they manage their assets and it goes beyond real estate and facilities. Obviously we're all here to learn about real estate and facilities, but it's for all asset classes.
Kal Gyimesi: Yeah. So the entire life cycle of everything that they have to do to be able to run their real estate operations and their facility operations, like a system.
Everything from acquisition all the way through disposal. And when you think about all the steps that go in there from acquisition to building out facilities, renovating 'em, you may be constructing them, populating 'em with the the data that you need, administering that real estate from month to month whether that's through sustainability or doing lease administration, all the space management and then maintenance, and then ultimately maybe disposal decisions down the line.
The, that inner circle may take a bit of a loop for a while. When we map out what our customers are using to do all that work, all of that work has to get done with any large enterprise that has dozens or hundreds or even [00:06:00] thousands of facilities that we work with. When we map all that out, what we often find is there's just a smorgasbord of different choices, that there's still a lot of silos, data silos, system silos, that break up that whole life cycle, and when you map it all out on a page it looks like a mess. That is very difficult to apply. Any kind of systematic decision making to,
Vishal Rane: yeah. So like some examples of that unstructured data, like could be organizations that are using descriptions for their work orders that they're assigning to go and do repairs or maintenance or replacements on.
That's a type of unstructured data because. It's not being mapped to an asset. Right? And then let's think about a KPI that you might want to move on in the future. Like total cost of ownership. Let's say on your capital planning side, you're using a different asset hierarchy than you're using on your operational side.
All of a sudden, that task of breeding [00:07:00] together your capital and operating. It's extremely complicated because it's not thought out for the end KPIs that you want to produce, right? Sometimes we're stuck living in the now rather than thinking about tomorrow. So that's one quick example, but when we think about it, what are we talking about?
We're talking about data governance, right? And what is data governance that's building your hierarchies? That's building your integrity plans for your data and then supporting that with the policies that you put in place within your organization. So sometimes it takes a corporate entity on a facilities or asset management side to bring together that thought leadership to get the organization to move forward.
And that underlying framework is how you're gonna build on ai, right? Because if the data inputs that you have to start with are not good you can't build on a strong foundation. Right.
Host: Properly supplying your business can be [00:08:00] stressful. ODP Business Solutions is here to help. They have all your break room and cleaning supplies in one place. From trusted cleaning products to healthier snacks, they have what you need to keep your spaces spotless and employees productive, and your operation running smoothly.
Whether you're in healthcare, hospitality, professional cleaning, and more, turn to ODP. Stock up your business stress free. Visit odp business.com to learn more.
Kal Gyimesi: You know, we still do a lot of manufacturing. About 50% of our facilities are the traditional office space and such that corporate real estate typically manages. But the other 50% is labs, manufacturing facilities. We still manage large computer systems. We are now building out quantum manufacturing, quantum computing manufacturing, which is required entirely new facilities to manage.
We're still doing chip manufacturing, so we have water treatment centers and all that. So our [00:09:00] corporate real estate teams have to, you know, they function like a large enterprise, and they've just gone through a large transformation over the last two years where they had about 48 different systems that they were at a complete stranglehold of being able to process information much less, start to apply AI on top of that. So they've had to go through a big digital transformation. And we've gone from about 48 systems down to about nine and that transformation has just opened up the ability to analyze and apply ai throughout our different processes and that's been a boon for us and that's gone a lot of the ways down toward achieving that $4.5 billion savings. One of the things that visual touched upon is the importance of data governance.
That's one of the things that they absolutely invested very heavily in. And they put strong process management [00:10:00] capabilities into place.
So the organization and data stewardship, having data owners and very strict rules on how, how the data definitions had to be created. They could only be changed through a certain process and all the metadata that went into the systems all had to be managed very carefully. So, when we're consolidating systems and putting good data governance behind it, now we can start to apply AI and add AI as an addition into workflows that we can make more efficient. What we don't want to do is plus AI take a mess of systems and think that we can solve problems by throwing AI on top of it.
Vishal Rane: Yeah. So really AI is about, you know, finding efficiencies and basically supercharging your workflows, right.
Kal Gyimesi: Yeah, so 72% of executives believe that you've gotta have proprietary data your own proprietary data in the learning [00:11:00] of the AI systems that you're trying to apply in order to be successful.
You know, build a modern data architecture that breaks these silos and gets enterprise data in a point where you can apply AI to them, co-create an AI governance framework. Very important to have good data governance and to not overlook that. Then we want to operationalize AI wherever we can enhance experiences and unlock new value.
Very important to test, retest, and keep testing again. Because AI will give you sometimes unpredictable results. So we want to be able to keep testing and really understand the processes that we're putting it into and what kind of results it's giving you back out. And then be vigilant to spot risks bubbling under the surface.
You can't just put AI to a process, walk away from it. And turn off other human capabilities. We've gotta keep testing it and looking for issues. It's a brand new technology. We're just [00:12:00] starting to put it into work processes and we have to keep testing and looking for the faults and while we perfect the way that it's going to work.
Host: Thanks for tuning into the Connected FM podcast. If you enjoyed today's episode, please take a moment to rate and review the show because it really helps us reach more listeners just like you. And don't forget to hit the subscribe button so you never miss an episode. See you next time.