InfrastructureAI Advances Autonomous Building Operations

ai Leadership Insight Series. Today, I'm with Dilip Rahulan and Glenn Almandinger, who are from InfrastructureAI, and we're having a little chat about autonomous buildings. We have autonomous cars now, but apparently, the buildings themselves could be a lot more autonomous, and I'm going to let them explain.

Dilip, Glenn, welcome to the show. Thank you. Thank you.

So walk us through what's going on here. We've had smart buildings for a while, so what's an autonomous building, Dilip? Thank you, Mike.

Thanks for having me. The concept of autonomous buildings are a traditionally radical approach towards ensuring that buildings are able to manage, optimize, and operate to expectations by themselves. And fundamentally, it's taking 100 years of technology development in infrastructure and transforming that into the modern era, facilitating artificial intelligence to play a significant role in ensuring that the building is operated, managed, and its energy consumptions are optimized using agentic technology that facilitates real-time applications and bringing intelligence to every device within the building and its ecosystem.

Glenn, how will this all come about? Because the current state of the art, at least in most people's minds, is that it takes a massive data center to deploy AI, and seems like you guys are talking about embedding AI agents everywhere across a building. So what will enable that to happen?

Well, essentially, partnerships with the silicon players and the equipment OEMs that allow us to leverage their hardware, processing it primarily at the edge. And then often in these operational technology environments as opposed to IT, OT versus IT environments, you have lots of people that want to manage this stuff locally, and obviously, this can require some adjustments and some evolution of the hardware base that feeds all this. But buildings and transportation systems, broad venues, airports, ports, things like that, these are all fairly operationally smart these days, just like your comment about buildings being smart for some time, but they're not very intelligent.

And I think in many respects, think of AI as an overlay to these existing systems that gets us out of operators sitting in front of screens and having to be reactive and essentially passive and reactive participants in how these systems get managed. And putting now essentially a whole layer, a new layer of AI and more meta automation into how these systems are organized, how they're orchestrated, how they get optimized, and how they're run. But in a distributed computing sense, you need to have tools at the edge that can process a lot of this much more locally and in a distributed fashion, and with a lot more powerful silicon driving it.

So Dilip, walk us through what will be the experience that will be different for people as they walk through those buildings, and what will change for the people who own those buildings? Fundamentally, the whole concept of IT infrastructure within buildings are fairly new. Convergence took place in year 2000, and since then, there's been totally a whole approach towards ensuring that every building has got its own network and is able to seamlessly function as a single entity in terms of its key components.

But if you look at the landscape of equipments within a building, it's quite disparate in terms of heating, ventilation, air conditioning as primary energy consumer and operational challenge functionality. Then you have everything related to horizontal and vertical transportation within the building. Then you have security.

Then you have key components of lighting and everything related to safety, all aspects of the wet and dry services that create human comfort. All these key elements within the building are now on a network facilitating seamless integration through automation. Now, what really happens is it's a programmed environment.

In an IT world, you talk about programming as functionally a methodology that enables things to talk to each other and facilitate adaption and working through a workflow, a process that would enable it to perform optimally. Now, once you program a building or once you commission a building from its inception, the whole process of programming and facilitating the building to run is based on what has been put in, set in, in phase one. And fundamentally, that process is not something that is dynamic.

You program a building and it drifts away in the first month of its activity in terms of bringing the occupants into the building and settling them down to their workspaces. The whole equation of a managed commissioning process is completely out of whack. So from that concept to looking at every element within the building, depending upon time of the day, time of the year, occupancy levels, weather conditions, all these matters influence a building like everything else.

And if you look at the dynamics that takes place, a programmed environment and smart building have- It is currently the order of the day, but an autonomous building is able to take real-time data, bring intelligence to every equipment that's within the framework of that environment, and make them operate dynamically. That changes the concept of having agents that can make decisions spontaneously depending upon time of the day, time of the year, occupancy levels, functional aspects that are determining the various forms that the building is being subjected to. And then there is the whole building envelope and its ecosystem, the whole dynamics of facilitating its existence in a city or environment where there are multiple buildings being serviced by service providers.

Fundamentally, the energy service provider, the wet and dry service provider, and everything else related to transportation and all functional aspects. Which in today's world, the building has no relationship with from a process that's being automated. But when you come to autonomous, you start becoming a part of the ecosystem.

You're becoming captively engaged in the dynamics of the service to the building and its neighboring buildings, its ability to correlate. So from a building owner's perspective, this is a radical transformation. It's inevitable.

It has to be done in order to bring him some very unique opportunities to reduce his operating cost, reduce his energy consumptions, and practically look at the building from a whole different perspective than what it is today. It becomes a living being, facilitating functionalities that it can do by itself. Glenn, what will it take to kind of build out this ecosystem here, and what are you guys trying to do and what's the role of the organization in all of this?

Well, I think my first order of response is that agents are only as good as the system they essentially reside within. And to Dilip's point about ecosystems and different suppliers, different service providers, different equipment brands, it's a mishmash and pretty heterogeneous environment that we're describing here. And we think at the core of this is essentially what we're calling an agentic operating system, and essentially an organizing principle, but also a set of standards for how agents can interact across different systems, different suppliers, different service providers, different stakeholders.

And if you don't have essentially an OS at the center of all this story, I think it's just a lot of competing solutions on a point basis and doesn't really yield the kind of end-to-end effect you'd want from that kind of solution that drives much more optimization of these systems. So even if you just take energy as an example, the HVAC systems and buildings in general consume about 40% of the energy on the Earth. Generally speaking, we think there's probably between 30% and 40% savings based on just optimizing time and use of that energy.

And so the ability to use agents to essentially orchestrate different elements of that system, and do so with the utility and the energy provider, and then the other sustainable resources, solar panels on the roof, that kind of thing. Then all that gets orchestrated, and the way these systems are organizing to utilize that energy aren't really just kind of an on/off kind of situation. They can be optimized and tuned and made to be, and to your earlier question, essentially a much more comfortable environment, much more tuned to the individuals using the environment within the building.

But also a lot less expensive and much more efficient for the owners and operators. Dilip, will these agents also communicate with the outside world? So if something is happening in that building, will more people will know about it?

I think a lot of the smart buildings that I know about today, yeah, they're smart within their four walls, but will the building itself become part of the larger AI community, as it were? Absolutely, but I have to qualify my statement, Mike. Mike, everything that's happening in today's world in AI is all related to the IT infrastructure that's existent in this world from a disciplined outlook.

You can very confidently say that there are processes and procedures established for tens of years guiding IT to be formulated in the fraternity of discipline. There are certain functional aspects that are being defined, and people follow that. In the building space, the manufacturer and equipment suppliers actually predominantly bring devices in multiple protocols.

Multiple ways they communicate with each other is a challenge. Really, every manufacturer. And interestingly, if you look at the landscape of equipment manufacturers, there are only a handful of equipment manufacturers globally.

They come in different brands, but it's the same company producing it in Australia as well as in Argentina as well as in Dubai. So fundamentally, there is a monopoly in terms of how these major equipment manufacturers deliver equipment into building and building infrastructure overall. Whether it is for government, whether it's for private sector, they all come under certain brand names, and they all have their own pedigree.

They create methodologies within the framework of their equipment, communicate in certain protocols, and they have this secret sauce in terms of the software that runs the equipment. So it has been a challenge for the whole fraternity in creating smart buildings to get all these equipments to talk to each other and fundamentally operate. And now imagine you are now come to a totally different dimensional change where intelligence has come to play.

Everybody's talking about AI from an IT perspective. Transporting that into the infrastructure space is totally not a seamless process. It needs a total amount of transformation in terms of how we looked at how to integrate multiple equipment within the framework of infrastructure.

And fundamentally, our company's approach has been to address this from grassroot level. The whole approach of the Galxy agentic operating system developed by InfraAI is its capabilities in ensuring that we are able to give agents a nomenclature, which simply means that every agent that we produce in our agentic factory has a title deed inside a blockchain environment exclusively created for AI, agents that come from the factory. Because unless the agent has got their passport to function and operate within infrastructure, it's going to be a totally challenging environment.

You can't have companies producing agents in their own whims and fancy to facilitate applications within infrastructure. You need to follow a complete paradigm shift in facilitating agents who have identity and able to operate for security purposes and also for functional aspects. So the whole strategy of the operating system is to have the capability that any building owner who registers on this platform has, first of all, got the opportunity to get to become a part of a blockchain registry for building an infrastructure.

So every building becomes a part of this registry that then gives a title deed for every building, and automatically, with the help of the agent, the building owner or the operator or the facilities management company is able to spin a digital twin. He can do it in phases with the equipments and subsequently get the whole building in a digital form, and that digital form is a part of a registry that is updated spontaneously. So the whole concept of blockchain within the infrastructure is totally different from the blockchain that's in the IT space.

We are here dynamically updating the functionality within the building, and that creates completely new avenues in fintech and other applications related to our platform, which I will definitely want to share with you. But the security base of the agent having a nomenclature is a completely new evolution from our side. These two blockchain ledgers work together.

They are seamlessly integrated, and they're able to now look at its functionality from the building perspective and the agentic perspective, and then the whole concept of laying out these agents into infrastructure becomes disciplined. That whole process allows us to now create an opportunity for all OEM manufacturers to become a part of this whole initiative visions and use them with a title, enabling them to have a passport to function as an agent for a chiller, enabling them to have a passport to function as an air handling unit, or anything related to vertical transportation, horizontal transportation system. These are all evolutions that have taken place and the way our platform has been established.

So this then allows us to address the ecosystem. We are able to now have agents who have identity. This has never happened before in the IT world.

Today, you have every new company producing these agents, and subsequently, some of those companies don't make it. They fold, their agents are floating around. It's really a challenge if you try to do the same thing in infrastructure.

It will actually create chaos. It will actually bring down airports. It will bring down institutions that are of magnitude of security levels are high.

So we really now are looking at this from a whole different approach in facilitating agents with nomenclature and passports to function within infrastructure, and then the ecosystem. Simple question, but can we retrofit this into existing buildings, or are we going to have to rebuild everything all over again to accommodate all these? No, you can retrofit.

Again, I think whether you think of facilities or transportation systems, anything in the infrastructure world, all these systems essentially are on, I shouldn't say all, but the dominant majority of them are on networks now. That world's inching towards more sanity and standards and interconnect and protocols and so forth. But at the end of the day, they're all collecting a lot of data.

There's a lot of telemetry going on in all of these physical applications. And so the ability to essentially overlay a control plane, essentially to Dilip's points, use blockchain to register both assets and agents and have those things coordinated and then be able to orchestrate the use of essentially what we're calling subject matter expert kinds of agents. You now have an ecosystem architecture that can essentially overlay itself on these existing systems.

And so, most large, expensive systems in buildings today are run by building management systems. It's really just a matter of tapping into that world and essentially integrating the data and then moving this into a much, much more orchestrated, optimized, auditable, and trackable, if you will, from a governance standpoint, system. So Dilip, what do you want folks to do?

Is there some sort of organization you want them to join or become part of to join this discussion? Or what exactly is the to-do list look like? It's a great question, Mike.

Glenn and I have been in the fraternity for over 40 years. Glenn has been nomenclatured as the father of IoT, and has probably serviced more than 100 Fortune 500 companies in terms of consulting with them, and creating an awareness that what is the next dimension of connected buildings and building infrastructure rollouts in terms of the future. The process that we have adopted is very different from the traditional methodology.

We're not trying to sell anybody anything at this point of time. Our approach is to get building owners to register into our platform to facilitate this whole concept of agentic transformation in a concept where they could literally experience it even on a PC. Fundamentally, if they have a computer within the framework of their office and facilitate data that's actually coming in and stored and analyzed in some form or the other, our agentic applications can be downloaded as a freebie.

There is no cost to that whole initial process. We want everybody who's a building owner or a facilities management entity to try and become a part of this whole innovative process. And in order to ensure that we are in the process of allowing them to register with us, facilitate the concept of downloading applications in the form of a container that is prescribed for that particular building from a perspective of registration.

We determine what's the size of the building, its location, and then we are able to interact and determine a process where they can actually spin a digital twin literally while they register and become a part of our ecosystem. This digital twin will then enable us and them to have the opportunity to try and test the applications that are developed by the agent in real time. So if a building operator is now downloading our application, he's able to give information that relates to the building, and once the agent is able to identify the building or give feedback to him, he can then give component-specific information, which is a very simple process of delivering data in the form of what's in the building as a system, and then that spins a digital twin spontaneously.

Now, this digital twin is the gateway for him to experience applications. " And then the agent would give him an answer, and he could apply it onto the digital twin before he applies it onto the building. This is the standard process we are actually rolling out.

And fundamentally, this creates an opportunity for the acclimatization process not to go berserk and launch something into the building that's supplied by some third party as an agent to optimize a chiller. That would be actually not conducive because it has to work with the whole ecosystem and the process within the building. And we are also offering very unique strategic partnerships with OEMs to facilitate the concept of embedding agentic applications using our gateway modules.

These gateway modules are prescribed to have computing capability, which we hope to develop as a strategic tool that will sit within the network of every building and allow to interact with the key equipments within the building. So the OEM partners can then subsequently interact with these modules and create intelligence that would allow the data coming out of a regular equipment in a retrofit environment or in a commissioning environment where you have a brand new equipment shipped to a site where there is an airport being built. Then you have these thousands of equipment coming in with agentic applications embedded onto them.

They all can shake hands and make that commissioning experience so unique that within a month you could commission a whole airport when it takes 12 months today. So there are some very interesting opportunities we are creating for the market that through our years of experience in this fraternity, and also working with some of the largest projects in the world, we have developed certain concepts that are, we believe, unique in the way we approach things. Our whole business model will revolve around getting maximum registrations into the whole platform where we have data then being analyzed with adequate data sovereignty being provided to every individual building.

The process would be to create opportunities to allow building owners to have applications which are not traditionally consuming their data, but to become a part of our marketplace where agents will be able to prescribe in case of any failure or predictive failure, projected failure, a supply chain model where we would be able to identify three suppliers as well as three contractors who could be able to deliver a service in an environment that is totally different from current marketplaces, where agents decide and offer the service, and then the agents can make a call based on the decision made by the human in the loop. Or it can also be given a mandate where it has given guidelines today, $10,000 to be spent in any emergency cases. The agent makes the decision by themselves.

Now, to enable all this, we do have the blockchain now integrated into a concept of a cryptocurrency that would facilitate transactions to take place in the digital space, fundamentally only for infrastructure applications and its processes. All right. Well, folks, you heard it here.

The whole way we think about constructing and managing buildings is going to change one more time, and what's interesting is that it's a combination of technologies. Agentic AI meets blockchain, and the output of all of that is much greater than the sum of the parts, once again. AI Leadership Insight series.

You can find this episode and others on our website. We invite you to check those out, and until then, we'll see you next time.