Dianna Hart: And now we’re going to turn to today’s topic: the power of AI, machine learning, and algorithms in the facility management world. I want to welcome my guest expert Simone Sams from ENTOUCH. Simone has more than 20 years of experience in data research and analysis, which brings enterprise-level solutions to market. Welcome to the show, Simone.

Simone Samms: Thanks for having me, excited to be here.

Dianna Hart: Awesome! So, as we’re talking about the topic of AI, it occurs to me that there’s sort of an elephant in the room here. When we talk about AI and machine learning, for a lot of people, the subject does cause a bit of intimidation, especially in the workforce because sometimes technology displaces people, or it evolves so rapidly that it’s very difficult to change and adapt. AI is in the news almost every day. Even Congress is currently having meetings about it, large technology companies are discussing the proper application of AI, and there’s even discussion around regulating it right now. I just think we’re at a place where we might know some of its implications, but there’s far more that we don’t know than we actually do. So, this topic is very top of mind for everyone involved in the workplace right now. There are conferences devoted to the topic of AI and even the topic of AI in the built environment. So, before we go any further and declare a bit of confusion about what exactly we’re referring to when we think about AI, machine language, and some of these other acronyms, it would be great if you wouldn’t mind just taking a minute to define some of these terms so that as we move forward in our presentation today, we know exactly what they mean.

Simone Samms: Perfect, yeah. So, let’s talk about that. One of the common questions I get from my team at ENTOUCH is, what is the difference between AI and ML and how are they related to each other? So, let’s start off talking about that. AI is pretty much just what it sounds like – it’s the practice of getting machines to mimic human intelligence to perform tasks. Most of you have probably interacted with AI even if you don’t realize it. Voice assistants like Siri and Alexa are founded on AI technology, as are customer service chatbots, which I’m sure we’ve all interacted with at some point. Machine learning is a tool within the AI toolbox. Machine learning algorithms are designed to mimic how humans learn from experience. For example, just like how we get better at recognizing images like cats after seeing many cat pictures, ML algorithms can improve their ability to recognize patterns and make predictions after being exposed to relevant data. So, now, as we all know, there are IoT sensors, we’ve all got our iPhones attached to us, there are devices everywhere, and all of these devices are creating massive amounts of data. This unmanageable, huge volume, and complexity of this data, and by unmanageable I mean unmanageable by humans, is now being generated and has increased the potential of machine learning as well as the need for it. I was talking to one of my developers this morning, and just here at ENTOUCH, with the devices we have in the field, we process about three billion data points a day. That is nuts, and it really truly is unmanageable for a human to comb through all of that data and find the important patterns and things that we should be looking at. So, again, just to kind of bring that full circle, AI is the practice of getting machines to mimic human intelligence, and machine learning is the algorithms that we build to help those systems learn how to mimic things that humans can do. Machine learning is one of the key technologies that help us achieve AI, but AI also involves other techniques that you guys may have heard of, and we’re going to talk about throughout our time this morning – natural language processing, speech recognition, and even robotics. So, those are all sort of tools in the AI toolbox.

Dianna Hart: Okay, yeah, that completely makes sense. AI is sort of an overarching understanding of all of these technologies that are necessary in order to make AI work in many different applications.

So right now, since we’re talking specifically about the facility management space or applications of AI, what are some of the key challenges that you currently see facility managers in their day-to-day roles facing in terms of operations? And how do you see AI addressing some of these challenges directly?

Simone Samms: Yeah, so at InTouch, we deal with medium to large size multi-site retail operators, and those property managers and facility managers are dealing with a diverse and demanding set of tasks every day. This ranges from preventative maintenance, repairs, emergency response to situations, managing space optimization, tenant engagement, and budget constraints. You guys know all this stuff; you’re dealing with it every day. There are a lot of areas where AI and ML can help, even in day-to-day management of these things, and then really getting into things like moving to a proactive maintenance strategy versus reactive, how you manage your data. Much of the data that you guys are dealing with are in data silos. How do we bring all that data together and use AI to bring to the surface the things that you need to be looking at, so that you are not having to analyze all of that data? Giving you real-time visibility into system performance so being able to log into a dashboard and see exactly what all of your assets are doing without you having to dig for that. That’s just time you’re spending. So using AI to bubble the most important things up to the surface so that you can just glance at those things. And resource allocation – all of these things here that we can do with AI. In my world, it’s to give you, the facility managers, time back to do the things that you’re really good at: the creative problem solving, the strategic plans. Let AI and ML do the easy stuff, give you the data points you need to look at to go build those strategic plans, but not having you spend time on those types of things. I think one of the biggest areas, as I touched on, is shifting from reactive responses to proactive. Having an AI/ML model analyzing parts of your data allows those things to bubble up to the surface, to help highlight those things that need to be fixed and addressed versus you having to look at them. With the rise of sensors, there’s IoT sensors for everything, advanced software, we’re creating huge amounts of data as I mentioned. So, any tools that we can give you guys to put in place to help you analyze all that data is where AI really comes in to help facility managers day-to-day.

Dianna Hart: A follow-up question to that, in my mind, would be: the tools that are purchased or provided, do they continue to learn about specific applications, say specific buildings or facilities, after they are installed and are working for a period of time within an application? Do they sort of learn additional things about the nuances of a particular facility, whether it’s climate-related or weather-related or asset-related? Are they continuing to build their own algorithms, or are they able to be configured additionally for specific purposes?

Simone Samms: Yeah, so I think the great thing is once we get all the data in, like once you outfit your building or assets with sensors and we start to bring data into, for example, at ENTOUCH, into our EMS platform, that’s where we’re really able to go in and start fine-tuning. Whether it’s your heating and cooling strategies, or looking at bringing in data from how space is being utilized, or occupation data, to start adapting the way that we manage lighting or different things like that, or pre-cooling, based on all the data that we’re getting back. I think about these models as living, breathing organisms. They’re constantly being tuned, changed, and adjusted, and the more data points that we get, the better. So if you bring in more data points, like adding occupancy sensors or anything with indoor air quality, and we start bringing that data into our model, then we can continue to fine-tune it. And our strategies and how we manage the buildings just get even smarter. So the more data that we feed in there, similar to our brains, right? The more data that we give our brains to analyze, the better that strategy is going to become over time.

Dianna Hart: Okay, that makes sense.

Dianna Hart: It’s really interesting to think about AI, as it is underpinned by machine learning, as an evolving technology that supports people and doesn’t necessarily displace them. But we have to adapt to these technologies in order to stay relevant, also in our own careers and jobs, and in the way we deliver solutions to be more and more effective as time goes on. Because this isn’t… you can’t put the genie back in the bottle. Things are not going to go back to the way that they were.

Simone Samms: That’s very true. I heard somebody say this very early on when I started looking into this type of technology, and I’ve kind of adopted it, which is: I don’t think that AI is going to take my job, but I do think that somebody who uses AI to their benefit every day might take my job. So, what I want to do is look at all of the ways that AI can help me and not just, you know, we’re talking about this from a facilities, HVAC, asset perspective, but I think as employees we need to look at it as how else can I use AI in my day-to-day to make my life easier, to give me more space. I think we’re all in situations where we’re working for companies right now where we’re trying to do more with less. So, I’m always looking for things that can help me in my day-to-day to free up brain space. The tasks that are repetitive, that are my day-to-day, what tools can I use to automate some of those processes and make that easier so that I have more brain space to do the really fun, creative parts of my job that are actually what I really love. For me, some of that is how can I use AI? If I need to build a deck for something, can I use ChatGPT to build a structure for me that gets me 50% of the way there, where then I can fill in with all of my experience that I have across multiple different sectors to build that strategy, right? And AI is never going to be able to do all of that; they don’t have the same lived experience that we do. So, there are components of AI, there are parts of what we do as humans that AI will never be able to do. And so, we’ve got to think about that and say, okay, let me make time for the things that are my strengths and let’s use AI to help me with the things that tend to take me a long time, that I don’t love doing, that shouldn’t take me a long time. And I’ll give you an example really quickly of one that we used at ENTOUCH. We’re a small company, right, and so again, back with that doing more with less, we have to find ways to get really creative to support our customers. And one of the things we’re working on, an initiative, is building out how-to guides. I came across, via a newsletter that I get called “There’s an AI for That”, which is great. It comes out a couple of times a week, and it’s just ideas on different AI tools that are out there and how they can help you. And I actually found one, and after I started searching, by the way, you’ll be shocked, you’ll start to get all these ads in your Instagram for great AI tools that you might be able to use. And I started getting ads for a tool that helps to build how-to guides. And this tool lets you do a screenshot of the process that you want to build a how-to guide for. So, I can go into my product and say, “Here’s how you change a schedule on your thermostat.” We record that, and this tool then goes and builds, with screenshots, a how-to guide. It doesn’t mean we don’t have to go back and adjust some of the wording, put it on our template, things like that, but it’s getting our support team probably 80% of the way. And for a really small team, that means that we can put out more content for our customer much faster. So, those are the types of things that we’re looking for to support our businesses, to again be able to support what we’re doing.

Dianna Hart: Right, and what occurs to me is that from a career standpoint and from a business standpoint, everyone needs to focus on the things that AI can’t do or can’t do…

Simone Samms: Well, yep, and those are the things that we need to focus and develop on in our personal and professional development.

Dianna Hart: Yeah, absolutely, to stay ahead of that curve.

Dianna Hart: So, that’s a great segue to my next question. In terms of examples of how we hear about smart buildings and we hear about IoT and AI-powered sensors and how they’re being used to enhance facility management and maintenance processes. Can you give us some examples or talk about how these smart buildings and these sensors are being used by AI to work better for customers and better for maintenance teams and facility managers?

Simone Samms: Yeah, so AI sensors and IoT devices specifically can do all sorts of things in the built environment. They can monitor the health of equipment, occupancy levels, environmental conditions in real-time. For instance, smart thermostats can adjust heating and cooling based on occupancy, which helps to lead to energy savings. The more sensors we have, the more data we can get. So we talked about some of the things that they can monitor. This gives facility managers an unprecedented visibility into their properties’ condition, again at real-time. We can start to detect anomalies, identify potential issues, and make data-driven decisions. I think one of the biggest areas that we’re seeing improvements, and I talked about this a little bit up front, but moving from a reactive to a proactive maintenance strategy, is by using AI and sensors. Predictive maintenance, I think, is probably one that almost everybody that’s here has probably heard about predictive maintenance and what AI can do in that area. It can help improve the reliability and efficiency of critical equipment by transforming maintenance practices again from reactive to a scheduled approach or data-driven approach. There are a few different ways it can do that, but we’ve got models that can help with early fault detection. These algorithms can continuously monitor equipment in real-time, collect data from all these sensors. The algorithms detect anomalies and deviations from the normal conditions that we would typically see. These anomalies can help us indicate that there’s an impending failure of some sort. So by catching these early, maintenance teams can address these issues before they escalate into possibly breakdowns. This helps to reduce downtime and impact to your tenants. You can start doing that maintenance at certain times and days when you know it’s not going to impact tenants or people in your retail space. Also helping to bring the cost down, right? The cost of a scheduled visit is much less than an emergency visit. So all these models can help us look at that. They can get you guys on an optimized maintenance schedule. So again, looking at equipment condition and usage patterns, and other things. When we’ve got all these data factors and we’ve got a model that can look at all this, because again, a human cannot look at all of this data, we can start performing maintenance precisely when it’s needed, versus waiting too long or doing it at a time when it’s unnecessary. All of these things help to prolong the lifespan of the equipment. If the equipment is operating at the optimal level, you’re going to need less maintenance on that unit. Less maintenance on that unit leads to fewer capital expenditures on replacements and upgrades. And then, I think, because it’s the business that we’re in here at ENTOUCH, is also increase energy efficiency, which I think we’re all really aware of right now. How can we reduce our impact on the environment? And we know that well-maintained equipment operates more efficiently. So these AI systems can help identify these opportunities for not only the proactive maintenance to keep the equipment in top performance but also help us to optimize energy usage, which leads to energy savings.

Dianna Hart: Right, yeah, perfect. Totally makes sense. So, what role does data analytics and machine learning play in optimizing energy consumption?

Dianna Hart: I think that’s something that a lot of people are looking at AI to help them do, especially with so much legislation now coming out to monitor and measure energy consumption around environmental reporting requirements. I think that AI is definitely on task to play a really critical role there. So, can you talk about AI in terms of sustainability efforts within organizations, especially public organizations and those that have to report against those things?

Simone Samms: Yeah, definitely. So again, because of all the data that we’re getting back, we’re able to analyze energy consumption patterns, identify inefficiencies, and then build models that can help suggest optimal energy usage. For instance, AI can optimize HVAC systems to ensure optimal temperature and humidity levels, which help reduce consumption and improve indoor air quality. Having the data that an AI-equipped EMS tool provides helps you understand the performance of the building system. For example, in the summer, cooling demands in an office or a retail space can be the difference between setting a new costly demand peak charge and staying underneath that. So, if we know or our system knows, we’re in the process right now at InTouch, we’re building out a model to bring weather data into the model that we use when we’re looking at how we adjust and how we build strategies for the right heating and cooling within a building. I’ll use a summer example, but if we’re in a situation where it’s going to be very hot, like under a heat dome, what we can do is start pre-cooling that space when energy is less expensive. No human intervention needed; the building, the HVAC unit knows what to do. It gets that information from the EMS system that says, “Hey, we know it’s going to be hot today, we know because of past performance about how long it takes for this specific unit to get to that temperature,” so it’s all very data-driven. If a building has used a pre-cooling strategy in the past, the algorithm could even look at that to help improve how they do the pre-cooling. Again, looking at historic behavior on specific units to make decisions about when to start or even when to stop based on occupancy and vacancy schedules in the evening. So, those are types of things that models can help us do versus a human having to sit there each day and look at “okay, what’s happening today, what strategy do I need to set.” We can let the building do that. With our tool, what we’re doing is we’re always going to have a little bit of human oversight. We’ve built tools into our portal. When this model is released, we’re going to come and tell you, “Hey, this is what we’re seeing over the next couple of days. We’re seeing this rise in temperature. Here’s a strategy that we would like to employ,” and you can accept or decline it. So again, there’s where that human element still comes in, but all the hard work of analyzing it and figuring out what the building needs to do, we’ve got the models to do that for us. So I think that’s a really good example.

Another example would be space utilization or occupancy. In a theater, we can start to pull in ticket sales, occupancy, and see what’s happening for that day of showings. Do we need to pre-cool that theater? Is it a matinee on a Wednesday with only five people in the theater? We might not cool that theater as much. So not only is your guest going to be more comfortable, but you’re also going to save energy and ultimately reduce that environmental footprint. All of these things go hand in hand as far as reducing energy expenditure, optimizing strategies, helping our clients hit their sustainability goals.

Dianna Hart: Yeah, so do the assets themselves, the HVAC systems and other assets that would be involved in the management of the environmental experience for everyone, do they have to be especially wired in any way in order for this to work? Or do they have to be new and especially configured for this to work? Or can older systems and systems that have been installed for a period of time also be available to be maintained and run by these artificial systems?

Simone Samms: Yeah, so old systems and new systems, there are different ways to talk to the systems. Here at ENTOUCH, we’ve got some protocols that we’ve put in place where we can actually work with our clients if they have older systems that are already in place. There are some communication protocols that we can use to communicate to those systems. Whether it’s new and you’re putting in new equipment that is specifically our equipment or somebody else’s, there are ways that we can communicate with that. It doesn’t always mean that as a facility manager, you have to do a rip and replace and that you need to go back and pull everything out and start over. There are absolutely ways that we can work with older systems, and companies like ENTOUCH can work with older systems.

 

 

Dianna Hart: Okay, okay, yeah. What about things like chatbots, virtual assistants, things like that? Can those improve communications and streamline requests?

Simone Samms: Yeah, so I think this is another area where we’re shifting away from just the HVAC and building management. There are other AI tools and ML tools that can help facility managers and all of us in our day-to-day business operations. So again, back to the idea that I want to free up my time and my customers’ time to focus on what’s most important to them, which is their customer, and where human interaction is required. Let’s look at some areas where we can use ML to help streamline some of those communications. You touched on chatbots. That’s a big one right now. Most of those use what we call natural language processing (NLP), so when you’re talking to them, it feels like there’s a human behind there talking. NLP is a huge area for the future, and personally, it’s something I’m really excited about and something we want to bring into our system more. We’ve got a lot of stuff built around automation, which is great – like when a work order comes in, the system should do this. But now, we need to start bringing in natural language processing.

For example, we had a customer enter a “too hot” work order recently. Our system, because it’s running just on automation, was like, “Oh, it’s too hot, lower the temp by one degree.” But when we went back and looked at it, in the work order, the person had said they heard a loud boom and everything stopped working. That’s where automation falls short – it doesn’t have the brain to process that additional information. That’s a perfect example where the added layer of natural language processing comes in and really starts to make our processes smarter.

The same kind of thing with chatbots. There’s a huge opportunity there. For those dealing with tenants, can you streamline communication through chatbots? Can we make it more personalized by layering AI into those chatbots? That’s a great example. Language support, multilingual communication using AI and chatbots – if you don’t have someone on the team who speaks Spanish or French, and you’re getting requests in foreign languages, that’s where AI can provide support.

The virtual assistant piece is really interesting. I haven’t played around with it much yet. I read in one of my newsletters this week, and a coworker and I, who always share AI tips and tricks, were discussing this. He sent me something on virtual assistants and how to create a schedule to help optimize time. Reading a lot of articles, it seems like everyone will have a virtual AI assistant at some point. I haven’t done a ton of research on that one, but I’m kind of excited to start playing around with it. I mean, who wouldn’t love to have an assistant, right?

Dianna Hart: So, well, yeah, here at Fexa, during most of our internal Zoom meetings, we do use AI note takers.

Simone Samms: Oh, yeah, we do too.

Dianna Hart: They are fantastic because they’ll summarize the meeting, and they’ll talk about what the outcomes are and who is assigned to what task, and email the notes to everyone.

Simone Samms: Those are, I would say, for companies, a really good tool to look into. It really allows you again to focus on the call and not be so worried about writing everything down and then be able to get that summary after the fact.

Simone Samms: Yeah, the other one I want to call out real quick, especially in the retail space, is the use of AI with video. Using that to see kind of comings and goings, using AI detection to look at suspicious activity. It can analyze behavior where you don’t have to have a human sitting there to alert you to things. A really cool area in retail is using AI in video to look at dwell and graze time. When you start talking about product placement in the retail environment, when you can start looking at how people are moving throughout the store and using AI to do that, it’s fascinating. Back in the day, I worked for a company where we would have to analyze footage of how people moved through stores. Now, being able to use cameras and AI to analyze how people are moving, what they’re looking at, what’s drawing their attention, that kind of stuff is really going to change the landscape of the retail environment. It makes all the sense in the world, fascinating.

Dianna Hart: So, with all of that in mind, what do you see in terms of looking into the future, the emerging AI technologies and trends that will have the greatest impact down the road on facility management? If you could look into your little crystal ball over there, what might you see?

Simone Samms: I think using AI for energy management is going to be huge in our space. Being able to bring in as many data points as we can get, I think, as sensors become easier to install and cheaper, anything we put a sensor on, we can measure. The more sensors we get out there, the more data we get back, the possibilities are infinite. It’s all about the data. I think that’s going to be huge. Using that to plan energy strategies for buildings, to reduce energy expenditures, and for companies to be able to hit their sustainability goals. As you mentioned, there’s a lot going on in that space right now around legislation. I think we’re all kind of waiting to see what’s going to come out of that. Employing AI strategies to help with that is going to be huge.

Simone Samms: Natural language processing, I’m really excited about, again, in our space, because I think there are some things that we’re using automation for, robotics for, but adding that next level of natural language processing is going to get us closer to that bit of the human side of things that right now some of the AI can’t distinguish between. And I think that’s really going to help us.

Simone Samms: AR and virtual reality, I know you and I spoke about that. I think this is still kind of, you know, there’s a lot going on with AR and VR and how that’s going to shake out. But I see ways that it can assist in facility planning, maintenance, training. Using AR glasses to access maintenance instructions and visualize equipment layouts, improving efficiency. I think there’s a lot of opportunity there that’s still to be explored.

Simone Samms: So, I think those are probably the three biggest ones that I see right now. I think there’s probably going to be stuff that we don’t even know about yet that’ll pop up as well. But those are the three biggest areas that I see AI still continuing to help us make big strides in this industry.

Dianna Hart: Yeah, makes sense. I mean, it’s exciting to think about what’s coming down the pike and how things are going to change. And it’s coming so quickly that it’s hard to foresee. And I think the key for all of us is to be open and to be adaptable and to be always learning. We just can’t stop being students and self-educating.

Simone Samms: Yeah, absolutely.. And I think I would say to anybody who, you know, as you start dabbling in it, start playing around with chat GPT or Bard or whatever you’re, you know, your AI or chat tool of choices, right? And start seeing what it can do. I mean, you’d be shocked at the things that you can ask it. And, again, just to help you structure the way that you’re building presentations, the way that you’re explaining things to people, um, it does so much to help get us, like I said, 50 or 60% of the way to a task. And then we just get to apply our creativity and critical thinking on top of the structure that it provides. And then the other thing I would say is look for areas in your regular life, maybe not your work life too, where you’ve got tools that might be using AI. For me, my first introduction really personally was through, I’m an avid cyclist, was through a training tool that I use. And they use AI coaching. And I’m all about betas and I was like, “Sign me up. I want to see how you guys were doing this.” And I’ve watched how they’ve tuned their models and it’s really fascinating. So find things, you know, whatever you’re interested in, um, you know, there, you, you probably got tools and you’re on your phone or things that you’re doing that are using AI that you might not even know about. Like you said, get curious about what they’re doing and how they’re doing it, and how they’re using AI and think about how you can use that in your day-to-day, um, day-to-day life and work experience.

Dianna Hart (Host): 100%. And just to comment on chat GPT real quick, the thing I had to learn and am still learning is how to craft the prompts. Yeah, it’s, um, it’s useful, and that in and of itself is a little bit of a task that takes a little bit of education. Yeah, I found a tool. It literally helps you build prompts, build prompts, right? Yeah, so you can get what you need, but it’s powerful and a little bit scary, but absolutely fascinating. And we’ll see where the future takes us.