In today's episode, host Tom Ryckaert, a Board member for the IFMA Belgium Chapter, and Erik Jaspers, an IFMA Fellow and Board member of IFMA EMEA, discuss the impact of artificial intelligence on facility management. They highlight AI's role in improving data quality, the challenges of integrating AI in FM, and the differences between AI in business-to-consumer and business-to-business contexts. They also explore how AI can enhance smart buildings and the potential effects of AI on FM job roles.
In today's episode, host Tom Ryckaert, a Board member for the IFMA Belgium Chapter, and Erik Jaspers, an IFMA Fellow and Board member of IFMA EMEA, discuss the impact of artificial intelligence on facility management. They highlight AI's role in improving data quality, the challenges of integrating AI in FM, and the differences between AI in business-to-consumer and business-to-business contexts. They also explore how AI can enhance smart buildings and the potential effects of AI on FM job roles.
Erik Jaspers: [00:00:00] The emergence of AI is not so much going to replace people is going to add to capability to do your work better. And so it will change existing roles. It will facilitate existing roles to do things faster and better. But most of all there would be my advice to any FM. This is coming. It's not going away. It will gradually enter our business from the business to consumer to the business to business market will enter our field.
Embrace 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 Ithma.
org to get started. In today's episode, host Tom Ryckaert, a board [00:01:00] member for the IFMA Belgium chapter, and Erik Jaspers, an IFMA Fellow and board member of IFMA EMEA, discuss the impact of artificial intelligence on facility management. They highlight AI's role in improving data quality, the challenges of integrating AI and FM, and the differences between AI and business to consumer and business to business contexts. They also explore how AI can enhance smart buildings and the potential effects of AI on FM job roles. Now, let's get into it.
Tom Ryckaert: So, first of all thank you, Eric, for joining us at World Workplace Europe in Rotterdam. And we started yesterday on the first day with an interesting keynote about artificial intelligence. And knowing you you have a lot of experience in this topic and this domain, mainly in technology and innovation, but maybe first, Eric can you introduce yourself, please?
Erik Jaspers: Yeah, well, it's a pleasure to be here to start with. I've been working now for [00:02:00] Planon, which is a provider of software IWMS type of systems for 25 years. I've been basically working in the area of of product strategy and innovation, so AI is something we're of course dealing with today.
Next to that, I'm like you we are both in IFMA engaged in certain roles. My role is being part of the EMEA board of IFMA. So that's actually the link to that. And I think both of us are already engaged with the IT community of IFMA and in some extent to WE as well.
Tom Ryckaert: Yeah. Then you're my ideal guest, I think, for this topic.
So yeah, coming back to that. So, yeah, I'm Tom Ryckaert, of course. And as you said, we are both active in IFMA. So I'm a board member in IFMA Belgium. So in IFMA Belgium, and also I'm the technology and innovation lead. So if something happens in our great country of Belgium, most of the time I'm aware of that.
So yeah, let's start and delve into the topic. So maybe first thing to start, do you [00:03:00] have your own definition about what is artificial intelligence? So our listeners and viewers are on the right track to start in this conversation.
Erik Jaspers: It's actually a very good, but also difficult question. I would say the thing is for me artificial intelligence has for the majority to do with the type of computing we're now introducing into the world. It has been there for a long time in certain areas, of course, in some, for some years. But what you see now is the type of computing that we're now introducing with AI is something that resembles how our brain works. So they called it years ago, actually, the technology was defined in 1943 in the UK.
Okay. Where they looked at the neurons in our brain and basically mimicked that into a, in a computerized model. So you have inputs, some [00:04:00] kind of a processing and outputs that you can go into new neurons. So that's why they called it neural networks. At the time, we didn't have the computer power to basically run a neural network like that.
But now, as we have progressed in silicon technologies, now it's available. So now we can basically put in hardware, so to say the neuron system that we have in our own heads. And that's why they talk about learning like we do as humans. We learn from the books we read, we learn from the conversations we have.
And so we get feedback from others that learns us. And so we build patterns. And that's actually happening in AI. So you and I were both in the IT industry. So our products that we put into the market have been programmed like if this is happening, do this else, do something else. So you can basically take the listing and you understand how the system will operate with AI. It's this learning, which [00:05:00] is creating a kind of algorithm that will produce the results that you're aiming for, but you can't really read how it's let's say arriving at that conclusion. So that's really interesting because it can do.
Very interesting things already today, right?
Tom Ryckaert: Yeah. And I think, as you said, it's really the next level of computing because it's not just zero and once it's really the complex structure of the different levels in the neuro computing, of course, and understanding the patterns will help us a lot. So maybe we can make the link to data because yeah, if you want to work with AI data is really important and maybe, yeah, you know it or not, but we are all working a little bit for Google.
Because if you know, you go to a website or you want to do a subscription form, you got CAPTCHA. And I don't know if you know CAPTCHA. So we're training already our artificial intelligence programs to identify patterns. For example, where is the bus in the picture? And you have to click on the frames in the [00:06:00] picture where the bus is.
And this is to train the AI program to understand the patterns because it needs to be trained with a lot of data. Yeah. So having that said for data how do you see this for the facility management community, because a lot of data is created. What will be the main challenges and how would you start with it?
Erik Jaspers: Well, first of all, to your point, and you basically point at it already. There's different avenues in AI because whenever we're going to discuss what Google is doing or Apple is doing or whatever company today, you have to really understand, is this a business to consumer implementation or is this a business to business implementation?
And I think there's a big difference there because everyone, when we talk today about the use of AI and examples of the use of AI. You will find them dominantly, like the example that you just [00:07:00] gave, in the business to consumer domain. When we talk about applying AI for FM, we are in the business to business domain.
And what's the big difference there? Business to consumer lives on the data that you provide in the public domain. So over the internet, the kind of thing that you just mentioned and stuff. Yeah. When we in the, when we are in the business to business domain, we serve clients that want to keep their data private.
You don't want to share that because some of it is very strategic. It provides insights in how the operation or the company is operating and how successful and so on. So much of that is very constrained in terms of where to apply it. So you don't want to have that in the public domain. It needs to be private.
So. How's then learning going to take place? Because so what I see happening in, in [00:08:00] the world of AI today is that there are some manufacturers who are going to set up models and a model, an AI model is to explain is a kind of set up of AI that is making it applicable to, for instance, image recognition or understanding how a bill looks like or as one has very, can have very specific purposes.
So what I see happening in the industry today is that there are companies who are crafting these models using public data. Then the melt model gets available. Then you can basically put it into your business and use it on your data. Right? So that's one second. Very important. If you're putting it in towards your data, there's two elements.
What's the quality of the data used to learn the model? And do you have confidence that data was of high quality? Didn't have too much bias [00:09:00] in it or whatever kind of thing. Then you get the model and then you have to care that the data you're going to attach to it is reliable and of high quality as well.
So I would say for FM to go, let's go to FM and apply AI for FM purpose. So out of outside the public domain stuff, there's some very important things that we need to basically look at and we still have to learn I guess we're the whole my statement would be For pure FM and real estate applications.
We're at the mere beginning. It's not some something that okay, the whole world is already doing it vendors will tell you Okay, I have it available and so on you have to do very important things first. Make sure you understand that you have the right models You For the right purpose, there's a cost associated to it.
So AI is one of the most expensive operations you can run. If you go [00:10:00]to Microsoft or AWS and ask them for the price of running AI, it's high on energy, it's high on computing. So it's costly. So apart from the great things I can do with it, I have to think about the economics. What's the payback on this great technology or application I can do with AI? Does it pay for itself in the end, right?
Tom Ryckaert: So if I understand you correctly everything about data, I think it's an important task or skill that needs to be added to the facility manager that they have a clear. Point of view on what data management is that they know what they currently are collecting.
Yeah. And then it's a second question could be, okay, is AI maybe a technology that we can use?
Erik Jaspers: Yeah. I think you're touching on a very important point here because when you've selected your application and the model that you're going to need to basically build something on that and the [00:11:00] data that you need for it internally to operate on you go into this data quality and structuring.
You have this new profession, data scientists, right? So the question is that going to be in the FM portfolio? I don't think so. But you need to understand as an FM. What am I going against? If I'm going to consider an implication like this and do we have all the resources in home to basically execute on it?
And what do you need for that? Probably your I. T. information management departments have a conversation. So my advice to FM's would always be from day one. You need to collaborate with your I. T. people, with your I. T. staff and stuff. I already on identifying opportunity because the market, to be honest, we work, we both work for commercial parties in the industry.
What I expect the [00:12:00] coming years to happen certainly the coming year is a lot of companies who sell solutions and software to shout into the market by marketing that's already there. And , as an FM, you might feel stupid. I'm not doing it already. Right. I think it's not the case for FM.
There are of course, advanced implementation areas, but they are dominantly in the business to consumer market because the market's address is huge. But for FM, if you run 20 buildings and you want to do something of AI in 20 buildings, it's not crafted out yet. We have to learn a lot and the market, the development of technology is going fast, so there will be new tools who make it easier.
You know, and so on as the as time progresses. But what I would like to state for today, for implementation on AI, first go look at the business case, identify potentials, put it in a, an [00:13:00] economic framework. Yeah. Am I really going to turn back something of it? What's going to be the expected outcome of it before you go into a project?
Tom Ryckaert: Totally agree with you because it's having this helicopter view, as you just mentioned. All these different parameters need to be in place. And even then, when we were talking about the data and how you're going to cope with that, I think also the GDPR question is a thing you already tipped it a little bit because if you look at Gartner and the predictions about, yeah, we are going to create super apps.
What is a super app, but you're going to share a lot of information you as a person, but it's looks hurdle. You have to take in business environment, but in your private environment, you're already doing it for years because you have Netflix, Disney plus, whatever you're giving all your data available.
And they will say, this is something you like, and this is your match. Even if you like a good glass of wine and you, you take a picture of the wine is going to say, this is wine. You will like, yeah. So it's making a lot of recommendations. So is this [00:14:00] something we also going to see in in the technology and development and super apps that are going to use this data or will this be a difficult one in Europe?
Host: Are you ready to connect with FMs who live in your area, work in your industry, or share your interests? IFMA chapters, industry councils, and communities have opportunities for you to be involved as you want to be. Visit ifma. org to find your group.
Erik Jaspers: I think there's a very important point that you're touching on that. I would like to compare it with the emergence of social networks, right? When social networks hit the market, it was great because we can easily connect. You can make friends, you can share information and it was all great, right? Now, looking at the downside.
What happened actually is these bubbles, people who only talk to these people and don't get information about other people in a different bubble with a different opinion, with different view in life or whatever, [00:15:00] but they would not be connected anymore because the system would basically, based on your existing preferences, strengthen that interaction.
So no go, not go beyond. Now, if we're going to add AI. To this question is, will it aggravate that problem? Will it make it worse even than it is already today because now we can have a very more detailed level can look at your preferences, but his preference the way to steer interactions with others. You know, perhaps it's much more.
Preferable to basically go beyond your natural bubble. Like we knew in FM, we need to talk in FM. We need to talk with all stakeholders of all kinds of all, let's say backgrounds and stuff, because we serve and we touch their working life. So I think in, in, in a general sense, it's very contra productive for an FM [00:16:00] to basically be enclosed in a kind of bubble of self provided preference, because it will.
make that life worse. It will make the principle worse. So it will, the risk there is that it will basically limit you rather than strengthen you on the other side. And then you go more in the, and in, in let's say the public domain AI, AI has of course, this massive, interesting feature of helping you out in things that, that Take a lot of time of you personally that they can basically do in seconds.
Like I give you an example. We are using a Microsoft co pilot in our company and I was late for a meeting. So half an hour late. So I got into the meeting and I just asked co pilot what have they been talking about? What were the action points of discussed as far? I got this set of lines, so I could read it in within two minutes, actually was [00:17:00] up to date of the conversations taking place in the meeting when I entered, so I could basically just flow into the meeting and that kind of thing, but that's more public domain.
That's not FM, you know, and that's great applications. And my advice would be do it work with that kind of these kinds of systems. In the public domain because you're going to understand more what it's going to bring to you and what's not. 'cause in other areas, copilot sucks. They state well if you have a story, you can make a presentation out of it.
Just ask ai Well, it'll come. It's not there yet.
Tom Ryckaert: Yeah.
Erik Jaspers: Right.
Tom Ryckaert: No, I think we have to take a lot of steps and also. When you read books about artificial intelligence or you go to keynotes, it's always the same. Yeah. The age of AI is now the age of a toddler in a human's life. So the progression we still have to make is really enormous, I think.
But even then, again, if you look at facility management for years, we both come from technology background. We're shouting digital [00:18:00] twins, smart building. How do you see AI having an impact on that?
Erik Jaspers: Yeah, very nice question because I was always irritated when when we went into the area of IOT connected devices, and then it was a marketing function of the industry.
They started talking about smart buildings. So a connected building to be a smart building, that was the language, but in my mind, it never hit anything like, okay that's, it's truthfully so because there's nothing smart about being connected. It's merely IOT will bring you connect connectivity and understanding of effective behavior of things by data.
Now that we're going in this era of AI, now we have the chance to get it smart, intelligent. Because now this basic layer of trustful [00:19:00] data, because sensor data, basically system, OT system, provided data has its well defined characters and so on, so you can work on that. Right. If you have the right models and stuff, then you can basically put that to use.
If you then can do reasoning and advisory, not automation of responses, that's, I think that's a little bit far off because the technology is not there yet. But if you can have an understanding of this massive amount of data, because that's of course the great thing about computerized AI, a computer can oversee terabytes of data in a second.
While we take a day to read a book of 150 pages, right? So the time we need to consume information and that of a computer is yeah, you can't compare that actually. So that's what I really think with when we are plugging AI on these types of IOT systems that we deployed [00:20:00] in our buildings. Then we can get to levels of intelligence and smartness.
Tom Ryckaert: So if you then talk about again, data, of course, and the different dimensions you have, you say in the past, it was more yeah, we're approaching lunch, so I will use some words linked to that. So it was a spaghetti. All the data was mixed together. Then we were trying to get it more structured, get more information.
So you get a lasagna, the nice structures, but now we go, we want to go further and we want to create wisdom. That we can take decisions.
Erik Jaspers: Yeah. I think that's the ultimate goal. Of course, to have some level of wisdom. I don't know how far I AI will help us in what time frame. I'm not sure by that, of course, but I don't know.
What I would like to express as well, we've been doing some data management and data structuring, but the question is, and I guess it's not this structures that we now see in the FM industry on data today, they need [00:21:00] another. They need another restructuring to make it addressable for AI. So that's what data science is about.
It's about taking the existing, but restructure it in such a way that it's consumable to the models, the AI models, the brains, let's say setups that are going to basically derive from it and draw some conclusions or analysis out of it.
Tom Ryckaert: Okay. So still a lot of steps to take. Yes. It's not about connected devices, but again, Yes.
Getting it more connected and get insights into it.
Erik Jaspers: You know what the thing is, when we talk about the use of AI in, in FM, you have to take in mind that you're going to start a project that you're going to run for 30 years. It's not over tomorrow. Oh, it's there. It's done. No, you need, you will need to maintain it.
You know, there's going to be new data sources that you're going to put into the models and there's the models are going to change. There's going to be new tools. So we just starting there, right? Right. It's just [00:22:00]there. I hope, I really hope that as we progress in AI development, that a good part of the AI developments will be geared towards making it simple to adopt and not so much only the capability of the neural systems themselves to basically drive new conclusions or whatever, but also make it simpler, to adopt, which would be very important for our industry because we're not I.T. people primarily, right? That's true. And all AI people. Yep.
So.
So, yeah. You have
Tom Ryckaert: to be aware of it.
Then,
maybe jump to another parameter that's important. I always like to phrase it in a positive way. It's the love for talent. We already saw that. Suffer from a lack of resources in the, in our FM community for a long time.
And yeah, you can read the article that's negative AI is going to take away our jobs or the positive way AI is going to help us focusing on more complex and creative tasks. What's your opinion about [00:23:00] that?
Erik Jaspers: The emergence of AI is not so much going to replace people is going to basically add to capability to do your work better, certainly at first, it's probably going to make a certain class of activities, let's say we can automate it in some way or something. So that will disappear, but there will be new elements to it as well.
Right. And and so it will change existing roles. It will facilitate existing roles to do things faster and better. So a kind of productivity element to it, quality element to it, perhaps. But most of all there would be my advice to any FM. This is coming. It's not going away. It will gradually enter our business from the business to consumer to the business to business market will enter our field.
Embrace it, work with it, do it personally. Don't shy away from [00:24:00]it. Try to basically put it to use also in your private life, do it safely. Very important. Well, there will be a lot of things there that embrace it because then after you, you've been working with it in your personal life or in areas of the work, like in teams and stuff or whatever, Google workplace and so on.
Use it because then you're going to understand what it's going to do for you and not, and then you can make. The let's say the mental jump to your own business and say Oh, would there be perhaps some applications of AI that is really going to help the business forward, right?
Tom Ryckaert: Totally agree. I think you have to embrace it and not, if you don't know it, there is always the element of fear and I think that you have to take away.
And if you do do experiments with that in your personal life, I think it can help really a lot. Maybe to end up a last question, because it's always the same. Yeah. You have a nicer idea. You're off. The manager, but how can you make sure that you can get it at the management board or the board [00:25:00] meeting to discuss these AI topics and make sure that yeah, they can invest in it.
Right. Would you be what would be your advice for this?
Erik Jaspers: Don't do it alone. No. That would be the advice. So when you're looking around this field, look who's going to basically who's going to be your buddy. And that's not another FM guy. It's probably an information management guy or a production guy or whatever guy. When you think of a potential implementation, who are the stakeholders?
And in those, in that range of stakeholders, who are the individuals that can basically together with you put forward a kind of proposition that will get senior management attention and prepare your, prepare it. Well, don't just go blowing AI. No, that's not good enough. Prepare the case and do it meticulously.
And then just put it forward, plan for it. Don't just [00:26:00] execute a shout around. That would be my advice.
Tom Ryckaert: Thank you very much for this advice and sharing your insights.
Erik Jaspers: Well, it has been a pleasure. So, see you next time, I guess. See you next time. Eric.
Host: Thank you so much for listening. I hope you really enjoyed this episode. And as always, please don't forget to rate, review, and subscribe to the podcast for more incredible content.