OTIS is a global leader in the elevator industry, with products that move 2 billion people every day. In this episode of CXOTalk, we interview OTIS's CTO, Ezhil Nanjappan, about digital transformation and data at OTIS.
Digital Transformation with Data at OTIS Elevator
Chief Technology Officer
Otis Elevator is a global leader in the elevator industry, with products that move two billion people every day. With this kind of reach and impact, it's no surprise that OTIS is committed to digital transformation, and they're achieving success by combining IoT technology and data analytics with their existing business practices.
Otis has integrated edge computing technology to collect elevator performance data and improve their business processes with predictive maintenance. But they also recognized that digital transformation must be meaningful for their customers and their business.
In this episode of CXOTalk, we interview Otis' CTO, Ezhil Nanjappan, about digital transformation and data at Otis.
The conversation includes these topics:
- On the technology of elevators
- Data collection and analysis from elevators (edge computing devices)
- How does data analysis support business outcomes and digital transformation?
- How does data collection improve the customer experience at Otis?
- On predictive maintenance based on data and condition-based monitoring
- AI-based predictive maintenance improves the relationship between Otis and customers
- Does new data lead to new revenue streams?
- Digital transformation strategy and culture change at Otis
- How does Otis scale its workforce?
- Does the “close door” button on elevators actually work?
- How does Otis prevent elevators from uncontrolled falls?
- On the future of elevator technology
Ezhil Nanjappan is the Chief Technology Officer (CTO) for Otis. He is responsible for applying innovative technologies to transform existing business models, products, and services to support business growth.
He played a critical role in introducing the Otis ONE IoT service platform, helped develop the suite of custom smartphone mo-bile apps that collect and share performance data, and supported the deployment of more than 30,000 smart phones to get digital technology in the hands of Otis mechanics around the world.
Ezhil joined Otis in 2001, and focused his career on advancing global IT operations, mobile application development and service transformation. He is experienced in all things digital – from sales force automation and cloud transition to agile development and global program management. He has received 22 global patents.
Ezhil Nanjappan: We have our own roughly 2.1 million elevators in our portfolio and one-third of our elevators are connected with the help of these edge devices which can communicate to the controllers and perform the data extraction and share that through cloud.
On the technology of elevators
Michael Krigsman: Today, we're talking about IoT, digital transformation in relation to elevators. That's Ezhil Nanjappan, Chief Technology Officer of Otis.
Ezhil Nanjappan: Otis is the world's leading elevator/escalator manufacturing, installation, and service company. We maintain more than 2.1 million elevators worldwide and we move 2 billion people today. That means we safely move the equivalent of the entire world's population every four days.
Michael Krigsman: Give us some insight into the kinds of technologies that are embedded in elevators and escalators.
Ezhil Nanjappan: It's a 170-year-old company. There have been a ton of evolutions along the way.
If you think back 40 years, we launched the first remote elevator monitoring coming back where we have the electromechanical equipment, where we have the controllers, drives, and we are able to communicate that and get that information and monitor that remotely and diagnose certain service activities but not being onsite. As technology evolved, we are able to take the data, analyze the data, and able to predict insights, and then decide what are the areas of tasks we need to perform in the job site.
As we progress, the technology in the elevators evolved over a period of time, so we have elevators – which you've probably heard about – that's run traditionally with robe. We changed that into belts, and we were the number one company to be belts. It's Gen2, which we launched in 2000.
Then also, we evolved from the technology side in the electromechanical equipment, and we make it more into where we are able to remotely access and perform most of the activities. That's the evolution we are going through from the side of elevator technology.
Data collection and analysis from elevators (edge computing devices)
Michael Krigsman: Elevators are electromechanical devices and you are enabling them in order to share data, collect data. So, tell us about that.
Ezhil Nanjappan: From the electromechanical equipment, we need to get the data from different places. One from the controllers, two from the drivers, and three from the door performance.
We put in an edge computing device. We are able to aggregate the data from the controllers, drive logs, and the door performance. Along with that, we added sensors to capture the door vibrations, the door noise, and process the data in a central cloud environment, and take insights associated to that to share with our customers as well as toward internal personas like our call centers as well as our field technicians.
Michael Krigsman: It's very interesting to look at elevators as essentially being edge computing devices, as you were just describing.
Ezhil Nanjappan: That's right.
Michael Krigsman: Tell us about the kinds of data that you collect. You mentioned a few things, but dive in a little bit more to the data aspect.
Ezhil Nanjappan: Let's start with the controllers. In controllers, we capture certain data types which explains about the state of the elevator and also in which landing floor right now the elevator is, whether it is in a door open or a door closed state. Also, if we take the sensor sitting on the top of the door elevators, we are able to capture the noise and vibrations.
Along with that, the controllers maintain a certain type of logs. We call that drive logs, event logs.
We aggregate the data in the edge device and then we share that data in a certain frequency to the cloud environment. Then we store that in our data lake, process the data, and then make that available to our downstream applications.
Michael Krigsman: These are IoT devices then, so the elevators are really part of the Internet of Things. Is that a correct way to describe it?
Ezhil Nanjappan: We call it elevator as a service, but basically we connect our edge devices to our controllers. Then there is an edge device which transmits the data from the elevator to the cloud.
How does data analysis support business outcomes and digital transformation?
Michael Krigsman: From a business standpoint, what's the purpose of doing this? You mentioned transformation as one of the foundation elements of your work.
Ezhil Nanjappan: One, to drive our service growth. As I mentioned earlier, we have roughly 2.1 million units in our portfolio, so we want to grow our service portfolio.
Two is to increase our pricing.
Three being the customer loyalty while making Otis more competitive against our independent service provider.
Just to your point, at the end of 2021, about one-third of our global units were connected and we plan to accelerate the deployment over a period of time.
Michael Krigsman: How does the data make that leap to supporting the business goals?
Ezhil Nanjappan: Now I'm going to drive into the downstream applications. As we walk through the scenario from the controller or the electromechanical elements with the edge devices, we are able to transfer the data to cloud.
Now let's talk about the data there. As I mentioned earlier, we bring all the data and store that as part of our data lake. The data lake builds the foundation for actionable outputs for real-time performance information, proactive communication, as well as predictive insights.
Let's take a few personas. One, our external customers. What they care about is what's the state of their elevator, whether it's healthy or not, whether the performance of the elevator is as expected, and what role the previous maintenance tasks or activities performed. Was it on time? That's from a customer point of view.
Now, if you expand the customer to like a skyscraper or a university where you have a campus where instead of seeing each and every elevator's health, they can use our customer portal application. They can see the entire health of the campus, and then they are able to double-click and drill down to a specific elevator in the specific building. That's our external customer, customer persona.
Now let's take the second persona, which is our internal, which is call centers. We call this OTISLINE (the users internally) where they receive the call from a customer when there is an issue with the elevator or performance issues with the elevator.
With this information, what we bring to cloud with the edge devices, our customers and our OTISLINE users are able to see, real-time, what is the state of the elevator. For example, you were a customer for a building, and say it's not working. If you're on a call with our OTISLINE users, they can literally see in which floor the elevator was two minutes prior or five minutes prior. Also, they are able to see what kind of error types generated from the elevator.
This will enable them to perform two operations. One, they can troubleshoot by themselves with certain conditions and safety rules. Two, they can pass this information to our remote experts who are trained enough and skilled enough where they can remotely access the elevator and perform certain operations under safety conditions.
Then if you're unable to resolve the problem, that brings the fourth persona, which is our field technicians where they dispatch to our field technicians. But at the same time, using the technology, we built applications and apps for our field technicians. They can see from their phone (for their buildings which are associated to that) what is the health of the elevator, what is the issue in the elevator.
Based on that, they could prepare if there is a parts replacement required or if there are certain steps what they need to perform. Then the app will guide through the steps to address the issues.
All these downstream applications are supporting our key four personas: external customers, OTISLINE (or call center) users, remote experts, and our field technicians.
Michael Krigsman: How do you enable all of these elevators, many of which have been in service for decades probably?
Ezhil Nanjappan: We have around roughly 2.1 million elevators in our portfolio, so we have taken certain processes and steps by regions, by countries, specific. One-third of our elevators are connected with the help of these edge devices which can communicate to the controllers and perform the data extraction and share that to cloud.
The roadmap of what we have is based on the controller types and based on the deployment process. We will try to extend the portfolio of connected units over a period of time.
How does data collection improve the customer experience at Otis?
Michael Krigsman: What does this all do to the relationship that you have with your customers, enabling this kind of data?
Ezhil Nanjappan: The customers could be a building owner or a builder who is building skyscrapers, or there are residential elevators as having residential owners, or the customer types can be in the airport authority or subway stations – X, Y, Z.
If you take this, and if you put in the customer segmentations, there are customers at high-rise buildings. They care about information like either available real-time (either using their apps or the applications), but also they care about integrating the data into their building management system. We have solutions which allow them to extend our cloud-based API to integrate with the building management systems.
If you take the other set of customer segmentation – as I said, the residential owners – there might be tenants living there or it could be an office building where they care about the performance of the elevator where we want to show, real-time, the health and status to the building owner or the building manager.
If you take another set of customers, like a university campus, schools, the campus owner should have visibility on the state of the elevators. That's where we enable the solution using our customer portal solution where they can see the campus view and then they can go to a specific building to see the health of the elevator, or they can go to the specific unit or the elevator to see the health data.
That's how we are sharing the information to our customers.
On predictive maintenance based on data and condition-based monitoring
Michael Krigsman: You mentioned earlier predictive maintenance on the basis of that data. Can you tell us about that?
Ezhil Nanjappan: Two areas where we focus on it.
Well, probably, I'm sure we all did it. We tried to keep the door open for a certain period of time to move a piece of furniture. It's difficult to predict this kind of scenario over here so, over a period of time, the performance of the doors may get degraded.
If you are able to detect that data using sensors and bring the data to the cloud environment, which is our data lake, where we are able to process the historical data and the real-time data coming from the elevator, then we are able to perform certain predictive insights. Like, for example, if a technician has a scheduled visit to go and perform door maintenance. If the health of the door in a particular landing floor or, if it's four floors, all the floors look good, then there is an opportunity to optimize that maintenance task.
Whereas if a specific door (based on the predictive maintenance) you're able to see an issue there, then that can be an additional task for our field technicians to go and perform actions and take care of it. That's how we are using our data using predictive maintenance.
Michael Krigsman: Subscribe to our newsletter. Hit the subscribe button at the top of our website. Be sure to subscribe to our YouTube channel because we can notify you of upcoming live shows.
AI-based predictive maintenance improves the relationship between Otis and customers
This presence of the technology and then the data that results enables you to provide better customer service and really, in some ways, change the business model, the core relationship between Otis and its customers.
Ezhil Nanjappan: I'd like to share an example here of one fine morning in one of the hospitals. The elevator to the surgery room was shut down or it was not working.
Using this technology, what I outlined earlier, our technician was able to see that information. Then, in order to take proactive action there, he decided and he went to the job site. He was in the front of the building lobby talking to the front lobby person saying that the elevator going to the surgery theater room is not working.
It was a surprise for them to know, first of all, how we are able to find out remotely and, two, he's able to show (using the app) that this is the problem happening in that elevator, so we need to go and take care of it. It happens to be he went there able to solve the problem in a timely manner. This brings customer satisfaction, as well as the stickiness to the customer, and transparency to the customers like what we are doing in order to fix or maintain the elevators.
Michael Krigsman: The transformation inside Otis also must be significant because you have lots of repair technicians and they're used to doing things differently, right? They show up and they physically look at the electromechanical device. Now you're asking them to relate to this piece of equipment in a very different way.
Ezhil Nanjappan: One of the areas where you will see the most transformation for us is in our field professionals. We developed a group of business applications for our field technicians to perform certain operations in an optimum manner in the job site.
I would like to highlight one of the applications. It's called the Tune App. Basically, this app leverages the built-in sensors and detect the level of noise and vibration. With that data, we leverage artificial intelligence and machine learning to recommend specific actions to our field technicians.
For example, if you want to do a floor-to-floor test or a door cycle or a brake test, they're able to perform all of those things with the app on their devices. That's a huge transformation we've been through with our service technicians.
Michael Krigsman: It sounds like Otis had to go through, of course, many different transformations over time. Where do you see this technology going?
Ezhil Nanjappan: You may have seen or heard about the service robots helping with the security or making full deliverables. We are now partnering with the robot companies, and we are communicating directly with elevators, the electromechanical equipment, and then able to perform access controls autonomously. That's a big game-changer for us.
You'll be able to see that in elevators. And one fine day, maybe elevators holding a door for you to go through. And the robots are helping the elevators to hold the doors as well.
It'll be deployed in hospitals and universities in a lot more areas.
Michael Krigsman: I'm assuming that you must have or your team must include software designers, developers, of course, but also the folks who are developing the electromechanical aspects because, after all, elevators (despite being enabled with all this technology), they're still mechanical devices at the end of the day.
Ezhil Nanjappan: Yes, there are. We have our engineering team members. We partner together. We work very closely in the technology transformation where we extend, enable, or transform the electromechanical equipment along with the software development. It helps to blend together to achieve the end-to-end connectivity as well as transform our business.
Does new data lead to new revenue streams?
Michael Krigsman: We have an interesting question from Twitter from Arsalan Khan. Arsalan is a regular listener. He always asks the best questions. He asks this. He says, "Has your digital transformation effort led to new revenue streams?"
Here's another interesting aspect. "How did you convince the executives and the frontline folks that this massive change is good for everyone?"
Ezhil Nanjappan: It's a journey for us. As I said, the service business is strong for us and we have 2.1 million elevators in our portfolio.
The transformation journey is, as I said earlier, we want to be transparent to the customers performing insights. When we put together an approach and strategy, we look into the list of areas and then the customer feedback.
The key thing customers are looking for is uptime of their elevators, minimize their planned shutdown time, so we have taken that into consideration and then see the current state and then the current maintenance tasks of what we have. We combine that together, and then how we can enable technology end-to-end in order to achieve the customer requirements. Also, in order to be transparent to our customers as well. That's how we evolve.
Then while we are going through this, from a technology side, we want to make sure we scaled up the architecture in a manner which can support this from an end-to-end point of view. Right from an edge device or edge computing, can get the required information or the key information to support the need what customers are looking for as well as the other personas I outlined. Then bring the data to cloud to support the downstream applications.
When we create the strategy and the approach, we took this holistic view from a business point of view, customer expectations, as well as from the internal productivity and efficiency perspective. Then we created the structure how we can move forward. That's the approach we have taken and we're able to move forward with this.
Michael Krigsman: I think about that kind of change, and it's such a massive technology shift because I'm just imaging here's Otis building these large machines, very heavy machines, and now you're not only in the machine business but you're in absolutely the software business (building APIs and so forth), as well as the data business. The shift is just completely massive.
Ezhil Nanjappan: As we are able to bring the data into our data lake, we can make this data to perform insights. At the same time, we can make the data available to our customers. That's where we evolve the journey of cloud-based APIs. Then that enables a couple of things.
One, as I said, the high-end customer segment maybe looking for the information to integration into that building management system. So, we are able to have data and the API as a service available which will enable a new line of business for us. That's one area.
Two, as I mentioned about the robots where we are working with the partners. We can enable and extend the API services (what we have) to take it forward.
At the same time, with the insights and predictive maintenance, we are able to increase our productivity because before going to the job site, me as a technician, I can see what is happening in the elevator. I can decide, do I need a part to go and fix the problem or can I fix it with a list of steps what I need to perform from a parameter side.
We focused on (exactly what I said) the topline business growth. At the same time, how we can increase our internal productivity.
Digital transformation strategy and culture change at Otis
Michael Krigsman: There's real value to not just the customer but also from an efficiency standpoint to the service, to the front-line service technicians themselves, which then encourages them to make that adoption, to make that change.
Ezhil Nanjappan: Absolutely, yes.
Michael Krigsman: It seems like that's also a very important part of this. I know when I talk with other business leaders, the culture change aspect of digital transformation always seems to be the most challenging.
Ezhil Nanjappan: As you know, we went through this journey five years before we started with our service transformation journey with our field technicians. The approach we have taken there is it's not that we're trying to transform the entire end-to-end overnight.
We did a lot of ride-a-longs with our technicians. We were able to identify the areas of opportunities where we can improve upon. Then we wanted to build on an incremental phase.
In order to do that, from a technology side, we needed to make sure the architecture can scale up and support that.
That's the parallel activity with my team. We're able to implement that along with the business team where we took key areas how we can implement.
Then the development cycle we wanted to optimize as well, so we went through a process where it's a standard process now, pretty much. But you have a workshop for a couple of days, standardize the end-to-end steps what we need to perform on the job site, and then we build apps in a 12- to 14-week period of time and make it available for our champions team who can deploy it in the job site and get the input and feedback. Then we increment by adding additional enhancements.
That's the journey we went through. Then we are able to see the benefits over there from a productivity and the opportunities. Then by having the connected elevators, the Otis One program, we are able to support other personas, including our external customers.
Michael Krigsman: Can you tell us about the composition of your team? I'm so interested in this because, again, when talking with somebody who works at a software company, it's pretty clear what the team will need to consist of, at least generally. But Otis is so unusual in having both this very strong hardware as well as the software aspect.
Ezhil Nanjappan: From the software side, we are transforming into a standard of dev sec ops. When we started the transformation journey, we started with dev ops as a standard practice. Then there is a training and change management, which we went through internally, so we have business partners acting as product managers. Then we have a technical team acting as technical product managers.
We combine and work together. We go through the standard agile process over here. But I just want to highlight a couple of key points.
Last year, we touched almost 15,000 story points from a development point of view. The story point, as you know, a few story points combined together enable the feature sets from an agile development methodology. So, we are able to accelerate to that volume by establishing standard development methodology, common tech stack, and leveraging the cloud-native services in order to accelerate our development.
Michael Krigsman: Koustubh Bhattacharya says, "Elevators are mostly public utilities. How do you really measure the user experience of lifts using data?" That's a really interesting question.
Ezhil Nanjappan: By putting the edge devices connecting to our controllers, we are able to measure the performance of the elevator. Having sensors, we are able to measure the performance of the door at the landing floors and respective floors.
We do not capture any of the (from a privacy and requirements point of view) people movement and other things at this point of time, but we do track the information coming out of the drives, controls, and other things, and we are able to make it available to our customers.
Michael Krigsman: Correct me if I'm wrong. It sounds like it is a function of measuring the elevator performance metrics, things like are the doors closing rapidly, things like that. Is that correct?
Ezhil Nanjappan: Yes, and then if we take a particular customer segment, probably you might have noticed we started deploying a screen inside the elevator, so the passengers, the riders, they can see the content like any live news, music, songs, and other things. Also, it displays the position indicator of the elevator movement.
There are a lot of other initiatives we have in our roadmap from enhancing the user experience point of view beyond the aesthetics and the screen inside the car or inside the elevators. That's in our roadmap. We are going through that.
Michael Krigsman: This is from Matt Wood. Matt says he knows that Rolls-Royce engines have a control room monitoring the status of every jet engine globally from a central control center in the UK. They know which ones are at different stages of their flights, the maintenance stats, and so forth.
He says, "In a sense, elevators, lifts, are similar to jet engines, although not quite so life critical. Do you see such functionality for lifts, that kind of control room monitoring, the ability to drill down to the individual elevator?"
Ezhil Nanjappan: We do have our OTISLINE, our call centers, at the regional and the country level where they are able to see the portfolio of units deployed or installed in the particular country. In addition to that, our field technicians are assigned with a certain set of units, and they can see that in their mobile devices using the apps (what we have built).
From a control center point of view, at this point of time, we are leveraging the OTISLINE or the call center users where they can see the visibility of the elevators in the particular country but not like Rolls-Royce sitting in the UK and see the entire thing. We are not there yet.
How does Otis scale its workforce?
Michael Krigsman: Let's jump over to Twitter. Lisbeth Shaw has a question. She says, "There's scaling of the technology architecture. How do you scale the human part of the equation?"
Ezhil Nanjappan: On the scaling of the technology architecture, if you see it, it's running as 100% cloud-native applications in our data lake. We are able to scale up the architecture to support millions of messages coming from all these elevators in a day.
From a human part to it, like you and me, we are the riders or passengers of the elevator. We have received, based on our interviews and discussions with the customers, a lot of feedback on the things we can build up on and enhance.
Take an example if it's in a subway or a train station. We know at 4:00 p.m. there are four or five tracks, trains that are going to come and stop there, which means are we able to put in a logic where we can park all the elevators in a place where they're able to get into the elevators and reach their destination in the shortest wait period of time. That's the kind of things we are working on, putting in automatic logic, and then the machine learning on top of it to minimize the wait time.
From a human point of view, based on the density and the areas where we go to work on, that in some areas, if you take smart cities, there is traffic movement where we can capture. If it's a public place like malls and other things, we can get the data and the usage.
We know in the month of August the usage has gone down. So, we can see, okay, let's park a few elevators because the volume has gone down. Whereas, in the month of November and December, the use is going to be much higher around shopping and other things.
These are the kinds of things we are working on and has been deployed in a few areas. We are piloting and getting feedback. That's how we're doing it from a human part, but there are a lot of things we have in our pipeline as a roadmap, which we'll continue on this journey.
Michael Krigsman: I suppose once you have this body of data and you've created the technology that allows you to analyze it, then your ability to do product development is limited really by your imagination, by what your customers need or tell you.
Ezhil Nanjappan: Correct. Over a period of time, we are able to aggregate and process the data. It helps us to perform optimum insights on the usage, the number of runs.
We can sit here, and we can see in a building what's the usage and number of runs. Then also, we are able to see the usage over a period of time.
I just want to give you a few data points there. As you know, during COVID time, the usage in the public place has gone down. We see that.
Then, as you see towards the end of the pandemic, we see the usage going up across all areas. We see the growth rate growing from 5%, 7%, 10%, so on and so on.
Michael Krigsman: This is from Arsalan Khan. He asks another really, really good question here. He says, "How did you invite or convince the business folks to be collaborative when it comes to learning about innovations inside and outside the company, and then actually executing and implementing those innovations?"
Ezhil Nanjappan: We co-partner, we work together, the business. If you take the business, there are certain field team members, marketing, and sales. We are all together as one team along with DT and engineering.
When we created the strategy, when we tried to decide the list of functionalities and priorities where we want to implement, we combined together with the help of product management, and we defined the priorities. We set the expectation by implementing or enabling this future. This is the ROI or the benefits we can see or the transparency of data which we can share with the customers.
We go through that process and then we put that (from an agile development methodology point of view) into our program implement plan, and we take that. We perform that on a monthly basis, so that's how we work together as one team, and we all know what functionalities we are going to enable and when it is going to be available to pilot in the field.
Michael Krigsman: You have a fairly formalized methodology or approach for aligning technology innovation with business goals.
Ezhil Nanjappan: That's right. Every functionality is tied with the business priorities and business opportunities. When we align it, when we discuss on a particular feature or an enhancement, we align end-to-end across business team members, product management, on the technical team, those who are working on executing the functionality.
Does the “close door” button on elevators actually work?
Michael Krigsman: I have to ask you this. Does the close button actually do anything on an elevator? [Laughter]
Ezhil Nanjappan: No, it does not. It does close the elevator, but it is not—
You know sometimes you can see some locations where people – I mean it takes like less than three seconds. But then, obviously, sometimes people try to press multiple times in order to close it.
But there are certain conditions we put in place when to close it. That's due to certain safety conditions and the leveling floor in the particular landing floor as well.
How does Otis prevent elevators from uncontrolled falls?
Michael Krigsman: Since you're willing to share with us the deep secrets of elevators, how do you make sure elevators don't fall? [Laughter]
Ezhil Nanjappan: It's coming from Elisha Otis. One hundred and seventy years before was the first time he demonstrated in the Bronx. It was on a rope so, by cutting the rope, the elevator stopped. That's where the brakes and pulleys were established.
As we developed from there, as we went into the rope in the high-rise buildings, there are a lot of controls and brakes put in place in our system. Even if it travels at the speed of three meters per second to eight meters per second, the brakes and controls are put in place. Then that's how we are able to land on a specific floor, whatever the speed of the elevator is.
Then we extended and expanded that into a lot more, which you can see the landing. If you go into a high-rise building, even if you go up 80 floors or 90 floors, you won't see the vibrations and noise. The way in which we put the ropes and then the brakes associated to that and the controls associated to it are in such a manner that the riders and passengers won't see any difference that they're moving six meters per second or eight meters per second.
Michael Krigsman: You're constantly monitoring that set of data as well.
Ezhil Nanjappan: Yes. Now we've started populating the rope performance and other data. It is in our roadmap to add that, so it'll bring additional opportunities for us as well.
Michael Krigsman: Okay. We have another question from LinkedIn. This is again from Koustubh Bhattacharya. He raises an interesting point. He says, in his experience, the biggest problem with elevators is housekeeping, not cleaning them on time. Maybe it smells because they're damp lifts. He says he knows it's not an issue with the elevator manufacturer, but it ruins the overall elevator customer experience. And so, to maintain the brand image of Otis, do you look at those nontechnology aspects and try to think how to make the elevator a better place, a better environment?
Ezhil Nanjappan: There is no visibility to get that information. But now, with our sales supervisors, service supervisors, and our field technicians, they visit the job sites and customers in a regular frequency, in a certain interval.
During that time, there are certain things which we enable in an app format, which is to perform some level of audit. That audit includes safety, performance, cleanliness as well. With that, we are able to collect that information and process that information and see certain things are responsible for us to take care, certain things are responsible for our customers to take care.
We share that information in full transparency what we are going to do on the job site as well as what our customers are going to do on the job site. That has been deployed in progress as well. But if it is a specific customer, I'll get some more information, and then I will take a look into it and see how we can help over there.
Michael Krigsman: We have a really excellent question on LinkedIn from Mark Brewer. He is VP of Service Industries at IFS. I'll say, Mark, we had your CEO as a guest on CXOTalk. Just search for IFS on our website.
Here's his question. He says, "Where are you on your predictive maintenance journey? Are you using AI and machine learning today to accomplish this?"
Ezhil Nanjappan: With the data, what we are collecting from our edge devices as well as with the sensors (what we have on the door performance), we bring that, and then we have AI and ML. Where we are able to predict the door performance – I was answering to the other question – we are able to see the landing floor is in line with what we're looking from the standard tasks point of view or are there any tasks we need to perform based on the landing floor performance or the door performance.
That kind of a prediction, we are making it available to our field technicians. Then it triggers that as a task for them. It will allow them to make sure they include it as part of their next visit. That's where we are using our predictions.
Also, in addition to that, the key thing is we want to get the feedback because we need to have a closed-loop process in order to make sure the AI or ML model of what we built has been enhanced and then extended. We get the feedback from our technicians and insert that so that way we can make our machine learning (and then the models of what we developed) enhanced over a period of time in order to be precise from a precision and accuracy point of view as well. That's the journey we are going through internally.
Michael Krigsman: We have another question from Twitter. This is again from Lisbeth Shaw who says, "How much can the CTO control elevator system design given the fact that you are—" she's making the assumption "—integrating a large number of components and systems from outside vendors?"
Ezhil Nanjappan: We have our engineering team, and we work very closely together as one team. There are certain parts and other things, yes, that come from external vendors. But then when we put it together from a controller point of view (drives and other things), that is coming from our team.
From a DT or CTO perspective, we work closely with our engineering team members, the data types, the normalization of the data, what we need to bring, make that available in the edge device before we surface back to the cloud for the downstream applications. We work together as one team, so the level of involvement at the controller level, the drives and other things, are as one team. We work together.
On the future of elevator technology
Michael Krigsman: Any final thoughts on the future of elevators and where the digital transformation of Otis is headed?
Ezhil Nanjappan: From a future perspective, there are a few areas we touched upon. One about the robust building management integrations and, from a technology side, the areas where we are working on. One, as we discussed earlier, about the cloud API, and then also the data in the cloud enables us to do the predictive maintenance.
In addition to that, the cloud transformation is a key thing for us, which we are going through in our journey. At the same time, serverless architecture is the other area where we are focusing on.
In the future state, we could involve some advanced training methods of safety measures using VR, AR, so that'll help us to optimize and train from a safety point of view for our field technicians and other things. These are all the things we are evaluating from a technology side.
There's a lot going on, and I'd say it's a very interesting journey. I'm sure at some time in the future when we discuss, we'll share more about the amount of transformation what we performed at the time.
Michael Krigsman: All right. Well, with that, it's been a very fast conversation. A huge thank you to Ezhil Nanjappan. He is the chief technology officer of the Otis Elevator Company. Ezhil, really, thank you. I'm very grateful for your taking the time and sharing with us today.
Ezhil Nanjappan: Thank you so much, Michael. Thank you for having me here.
Michael Krigsman: A huge thank you to everybody who watched, especially to those folks who asked such excellent questions. You guys really are such a smart, intelligent, sophisticated audience. I love your questions. Always keep those questions coming.
Before you go, subscribe to our newsletter. Hit the subscribe button at the top of our website. Be sure to subscribe to our YouTube channel because we can notify you of upcoming live shows where you can again jump in and ask your questions. Thanks so much, everybody. I hope you have a great day and check out CXOTalk.com. We'll see you soon.
Published Date: Nov 04, 2022
Author: Michael Krigsman
Episode ID: 767