Engineers are primed to think on their feet and solve problems in record times, but dealing with the weight of a whole country’s emergency call service requires lightning approaches and a ‘nothing is impossible’ attitude.

Today we dive into the world of the most important telecommunications operation in the country, the Emergency Call Answering Service (ECAS). We learn how the service operates in ways people may never consider and the contingency plans that help it weather any storm, or pandemic.

Our expert guest has been at the cutting edge of data and communications in Ireland since the introduction of the internet and is now Head of Operations with ECAS, Michael Kelly.

Listen below or on your podcast player:


  • How Emergency Call Answering Service (ECAS) operates
  • Problem solving in an industry with 99.999% uptime
  • Adapting to weather and pandemic phenomena
  • Lessons learned from introducing the internet to Ireland
  • Why we shouldn’t fear AI and start seeing it as an asset

Michael Kelly - Head of Operations - Emergency Call Answering Service (ECAS)

Michael has extensive experience in the telecommunications industry including his role with PostGEM where he helped introduce the public internet during the late 80s/early 90s. He has also served as Director of the Internet Services Provider Association of Ireland until he joined BT as their Head of Engineering Planning & Design. Since 2012, he has been Head of Operations for the 112/999  Emergency Call Answering Service (ECAS).

Engineers Journal AMPLIFIED is produced by for Engineers Ireland.


For your convenience, we include an automated AI transcription

Dusty Rhodes  00:00

Right now on AMPLIFIED the Engineers Ireland podcast, we get behind the scenes at 999 and hear how their engineers handled the biggest emergency of our time.

Michael Kelly  00:09

We brainstormed on Thursday afternoon. We had a prototype on the Friday morning. We proved the prototype Friday night. We built our production officers on Saturday and Sunday. And the first calls were taken from home Monday afternoon.

Dusty Rhodes  00:26

Hi there, my name is Dusty Rhodes and you welcome to AMPLIFIED the Engineers Journal podcast. In this episode, we're about to dive into the world of telecommunications and hear how an engineering mindset is vital to keeping up with operations in a fast paced industry. Our guest today has worked extensively in the area where his career has taken him from the birth of the Internet in Ireland to the last few years, where he's acted as the head of operations of the emergency call answering service with BT. I'm delighted to welcome Michael Kelly. Hi, how are you doing?

Michael Kelly  01:02

I'm great. Dusty great to see you. Thanks for having me.

Dusty Rhodes  01:08

Listen, can I start with the emergency call answering service, it's kind of something that we take for granted, you just dial 112 or 999. But a few of us very few of us understand how it actually works. How do you explain the service to people simply? Well,

Michael Kelly  01:24

I think everybody's familiar with the concept of dial 909, or 112. And you're put through to the emergency services, I think the majority people probably assume that it's a guard, call taker or a guard who actually takes the call. That's not how it works. And in most countries, that's not how it works. Generally, there's what we call a stage one service. And that takes the emergency call and determines with you the caller, what is the emergency service that you really need. And in our case in Ireland, that can be Garda, ambulance, fire or Coast Guard. So that determination is made, we gather some information, we also are gathering some technical information in the background that you wouldn't be aware of. And then that is passed as a package, the video, the data, the metadata, to the emergency service, and that you get the help that that you need.

Dusty Rhodes  02:19

I have in my imagination that people are making a phone call, you talk about gathering information that people aren't aware of are you able to take things like you know, could the location of a person's mobile phone or their number or where the area they might even be? While

Michael Kelly  02:34

like most things in life, it's about location, location location. If we can find you, we can help you. If we can't find you, we can't help you. It's as simple as that. So the technology has been improved, I suppose from 99 goes back to about the 1930s, where location would have been communicated verbally. But the problem with that is that the caller may not be particularly variable, or they may be completely unaware of where they actually are, or they have unfortunately met with such an incident that they no longer remember where they are. So to get over that technology has become even more important than actually getting an address or location from the caller. So over the last probably seven or eight years, we've improved our mobile communications, particularly such that how it works now for the majority of calls and the majority of calls these days, probably 75 80% of emergency calls are made on a mobile phone as opposed to a fixed limit. In the background, the handset is using various location technology is giving a GPS coordinates of exactly where you are. And in parallel to the call, we use SMS, where there is no record on the phone at all of us. But it's used to transmit and transmit continuously updated locations. Typically, the first location that we get is relatively inaccurate, it might be maybe 200 meters accuracy, but by the time we get the second or third one, it may be down to two meters accuracy. And that happens within 15 to 20 seconds.

Dusty Rhodes  04:18

First problem I can see coming up with this is GDPR and data protection, they're held to how do you get over that problem? Very,

Michael Kelly  04:25

very simple. There is an actual carve out in the data protection legislation, which basically says in layman's language, well, if this is an emergency, then all bets are off, and it's in your best interest that we're able to find you. Now having said that, we go to enormous lengths to protect that information. People probably assumed that you know, we just pass on the location willy nilly. No we don't. It's only passed to the emergency services. Occasionally we will get requests for call recordings and other info about calls. And it is only when people have satisfied the very, very stringent, most stringent requirements that our call recording might be released. And of course, it has to be strictly relevant to the person themselves.

Dusty Rhodes  05:12

Giving people help and getting them help fast and knowing where they are. I mean, they're all very important of what kind of levels are you dealing with? I mean, how many calls you get a day or across a year? How quickly do you answer calls, that kind of stuff uptime is another?

Michael Kelly  05:25

Yeah, uptime is another thing. I'll come back to that. to your first question. We do 2.4 million calls. So 200,000 a month 50,000. A week 6000 A day. Having said that, that isn't a complete answer. Because it's, I suppose I would say that emergency calls are extremely predictable. We're dealing with human beings worldwide. And they work in very predictable patterns. And that's an area that I'm very interested myself to share predictability of it is is quite fascinating. But 6000 calls a day, it probably folds to maybe five and a half 1000. During the week. As the weekends come in, it gets a little bit busier. And in particular than Friday nights, and Saturday nights would be the busiest of all. And clearly, Friday nights and Saturday nights can also fall into the small hours of the following morning as well. It also has a there's a little bit of a different pattern between emergency calls to police, our guard and to ambulance ambulance, the volumes grow through the day on a very, very gradual basis, peaking probably around 11 or 12 o'clock at night, whereas Garda calls would probably start to peak earlier in the day will be much more erratic, up and down. The other thing that's fascinating as well about it is that a guard a call from start to finish, and one of them one of our jobs is that we record everything and all data for evidence purposes for the courts and for for investigations. The guard call typically takes about two minutes, 120 seconds, the average ambulance call takes about six minutes. But with a very, very long statistical tail, we would have some cards that would go up to 2425 minutes. This could be for a number of reasons. Either there's a difficulty with finding the person, or actually there is a paramedic providing information or instructions to the call or or the the victim Unfortunately, while the ambulance arrives. If I'd say one thing about e commerce, it's all about data. It's really, really very statistically driven. I'm

Dusty Rhodes  07:38

fascinated to hear how you say that the calls to emergency lines are very predictable. It's not anything I would have expected you to said, Can you give me an example of that kind of predictability on on a call? Well,

Michael Kelly  07:51

I can tell you that the this if you like what I would call the safest time of the week, is about 1030 on a Tuesday morning. And basically how I would rationalize it is that everybody has either gone to work, or they've gotten they're already in school, or haven't gotten out of bed yet. But they're not going anywhere. Because actually the chief determinant of nine on calls, believe it or not, is weather. If the weather is bad, we'll get more calls. Now, that doesn't mean that more people are necessarily out and about, it just means that they're more likely to get themselves into a spot of bother more likely to have a car accident, they're more likely to trip, they're more likely to slip. And obviously, the more severe the weather becomes, the more accidents that are likely to have. Ironically, even though there are probably fewer fewer people out in the boat. The other way of looking at it is that and I suppose unfortunately, and this is not unique to our but in most countries in the world, Friday nights and Saturday nights are where people get themselves into the most trouble, probably the most severe trouble. And it's it's when response times by the emergency services tend to backup a bit. Now, I would say that emergency services are built around the peace. But during busy periods, there's always going to be some sort of a of a wait. But the job of Ecosse is to somehow ease the path through to the emergency services. And even if they are busy, and they're not in a position to answer that particular emergency call that we will do a decision to reassure the caller make sure that they don't title because, again, in a panic I think a lot of people's instinct is I'll hang up and dial again. Well, if you do that, you'd go back to the beginning of the of the queue. So our advice is always just stick with us. Listen to the instructions. We'll get you there.

Dusty Rhodes  09:51

Can I ask you that as well about uptime? It's a phenomenal statistic. What is it? What is your uptime guarantee? Our uptime

Michael Kelly  09:58

guarantee as well, first and foremost, some people say, Oh, well, it must be 100%. Well, as an engineer, I know that nothing is ever 100%. So what we commit to contractually, and I don't think this is a state secret in our contract with government, but it is what we call five nines, 99.999%. Now, that's a very glib figure. But the truth is to make that work, we have to duplicate replicate quadruplicate systems, so that we've got a huge amount of redundancy in the various systems. Basically, no one issue can take out the entire platform. But the bigger challenge for us is that the system itself, we we, we need to maintain it, we need to patch software, we need to replace hardware, but it's it's like the It's like that old adage about the 747 in the air, we're changing the engines without landing the plane, we cannot say to the public, oh, we need to do a big job on E casts. So I'll tell you what, we're going to take it down nine o'clock on Friday, but we'll be back on Monday morning. That doesn't work. So there's never a good time for us to do maintenance. So therefore we do, we're constantly working on the system round the clock, and making sure that our change control is absolutely state of the art engineering was so that, even if we do make a mistake, or if we have a problem, we can roll back without anybody realizing that there was ever a problem in the first place.

Dusty Rhodes  11:31

It sounds like you have your system and then the backup for the system and then a backup for the backup system. And then a backup backup for the backup system, which you know, I'm delighted to hear that but from an engineering perspective, can you give me an example of the kind of infrastructure that you have in place? Okay,

Michael Kelly  11:47

yes, we have two operator centers. So this is where call takers can take calls. We also have connections to them, they can also work from home, which is also part of our contingency in case, we have big storms or something like that. We have two data centers, we have to backup operator centers. And we have to backup data centers, all interlinked using multiple carriers by multiple telecoms companies. Some people think one, once they hear that BT operate this, that it's all BT telecommunications links in between all of the sites. And with that number of sites, it gets incredibly complex to make sure that we've got redundant paths, and resilient links and so on. No, we actually use every telecoms provider in the state. So we use ESB telecoms, we use IE Nash, we use air, we use BT in all honesty, and a couple of other players as well. And we use a variety of even within those telecommunications networks, a variety of different telecommunications protocols and techniques, so that we're not reliant on just one protocol, like IP or something like that. We have we have backups, little backups.

Dusty Rhodes  13:06

And tell me about the engineers that are in the organization because engineers play a very important role within the cancer organization, what kind of problems that they have to solve on on a regular basis, I

Michael Kelly  13:17

suppose the basis would be that they would be all IT specialists with a very heavy emphasis on telecommunications as well. But on top of that layered in, they would have skills in in software, but also a very, very good working knowledge of handsets, particularly mobile phone headsets. And you must remember as well that we've got to be able to support calls the highest level with the highest bandwidth that we could get from the humblest, oldest Nokia phone that somebody only uses once or twice a year, all the way up to the latest Apple and Samsung handsets. The other thing that we need to watch out for is software changes on the handsets, sometimes inadvertently, depending on the manufacturer. And I won't mention any names. But sometimes issues creep in with regard to emergency calls, that were actually designed to help the more ordinary run of the mill, cause a good example of a of this year was it was in the, let's say, the Android sphere. But it had an impact on all manufacturers of Android handset. So it wasn't a particular manufacturer, a change was introduced. So that if you picked up your Android handset irrespective of make if you pressed a number of times on the side and I don't want to specify the number for obvious reasons. If she pressed a button on the side of the phone a number of times it would automatically make an emergency call. Now that generated over about a year. I don't want to put a finger on it either but a shoe huge number of silent calls, calls that should never have been made. Because people didn't know when they might notice that that evening when they pick up their phone, and they see all these 112 calls, and that that was just basically down to a software change that was designed to help other issues within within the phone. As emergency services, we don't want it to be too easy to make emergency calls, we want it to be delivered. Okay? Because otherwise, you can have a situation where, you know, people literally walking down the road with the phone in the back of their jeans in their back pocket. It can it can ring emergency services. So it was well intentioned, but it went wrong. I suspect that problem generated probably a billion calls worldwide. Wow, it would have affected every country. Well, what I can say is because because of our engineers, I'm one in particular, I think we were probably the first country to identify what the problem was. And then in conjunction with, say, our colleagues and other in other European countries, and we do work very, very closely together. And because, as I say, dealing with human beings who do tend to behave in the same way, they use the same sorts of equipment to interact with emergency services. If we see a problem in one country, 10 to one, you're gonna see it in every other country. So we do cooperate very, very closely. So that's probably the best recent example that I can give you. I

Dusty Rhodes  16:28

don't want to dwell on this because it's over, fingers crossed. But when COVID hit, that was an emergency that was developing so fast, and everybody just had to run with it. I'm sure E. CASS was no exception. And everybody was told to work from home. How did you handle that problem?

Michael Kelly  16:44

Well, right up to that point, that was the middle of March and in that year, and I can't remember which year it is now, because it's all a blur. But up to that point we didn't have working from though there was never any requirement for us. And in fact, if I'm honest, Our preference would be to have people working in centers, because all the technology is there. And they have access to engineers. They have it. They're working in centers, which are designed to work 24/7 have generators, if there's a power outage, and so on, which is completely unlike our own homes. So I think it was a Thursday. And we said right, we need to build a remote working as solution. So we brainstormed on Thursday afternoon, we had a prototype on the Friday morning. We proved the prototype Friday night. And we built it or productionize it on Saturday and Sunday. And the first calls were taken from home live calls, not test calls, Monday afternoon. Now, I will be the first one to say that, you know, we work in in BT, which is, you know, really as an engineering lead company, I think if I had put out a request, let's develop that capability in normal time, it probably would have taken six months, because we would have gone through all of the things that you have to do in terms of testing. And, and so I'm not to say that we didn't do all that we just did a hell of a lot quicker, which much, much more focus. But we then moved over to a situation where very, very quickly, I think probably 70% of calls were taken from home. The other thing that we had to do, of course, is that staff were used to working with a workstation rather than a laptop, so would have had a purpose built PC, essentially in old money on the desk in front of them. So we had to we've we've about 6570 people taking calls around 24/7. So we had to procure 70 laptops plus spares at a time when everybody else was looking for laptops. Now luckily, I was able to pull in a few favors and BJs a big company. So we were able to get you know, access to certain stocks and so on. But that was the other worry, you know, it's one thing to get the technology right. But then are you actually will you have the tools to use. And I think that's probably where a lot of companies organizations probably slipped up well meaning in a well meaning manner, they were able to get the technology to do what they want, but it just kind of fell off the last part.

Dusty Rhodes  19:18

That was an amazing feat. And all I can think of is as you tell that story is this is why we need engineers in the world. Boom, boom, boom, problem solved. Tell me aside from COVID, because that is an exceptional circumstance. What would you say was the was the second biggest emergency that you've had to deal with in your time?

Michael Kelly  19:35

Well, I did say before that the biggest determinant of emergency calls, particularly when it gets out of that predictability phase is weather. And I will think back to the various storms that we've had on this one back in 2016. We've had some years where we've had maybe one or two kind of hurricane type storms and then over years we We've had maybe nine or 10 success weekends, where we've had really, really bad storms, that has an effect on the public because that, you know, there's going to be more accidents, even if they don't leave the house, they might fall down the stairs, and that that still creates challenges. But what it meant was the weather was so bad on these occasions. And don't forget that we had not got home working, or remote working available at that stage, we still needed to get people in sites. So what we did was we increased the number of sites. And that's when we, in addition to our two permanent sites, we brought in we built contingency centers, so that the centers were actually closer to staff. And then the other thing was, with great cooperation from our staff, I have to say, we, we said, Well, look, we might, we may have to work longer shifts, but we will put you up, we may even put you up in in centers, right with sleeping bags. Now actually, as it turned out, we didn't we didn't really have to go that far. But what we did do was we would put staff up in hotels next door, so that they didn't have to go home. And our only ask of them was look, bring a bag, we'll bring it back for, you know, five days or a week, we really don't know how long this is gonna go on. But we will, you know, a bit of, I suppose bit of thought goes into us while it's stressful at the time, and you think, Oh, how are we going to solve this problem? Nothing is impossible, if you set your mind to it. And that, you know, genuinely is our mantra is has to be that way. We can't just give up. So I think I think that would probably be the the other the other issue. Obviously, we've had technical challenges over the years. But generally, through a combination of backups, and so on, we've been able to overcome that. And the great Irish public wouldn't have even been aware, Michael,

Dusty Rhodes  21:51

you're very much at the cutting edge. If you want of technology. I hate that phrase. But you're dealing with Watson now, which is fantastic. And you're probably looking at what's coming in the near future, which is fantastic. Let me take you back, though. To many, many moons ago, when you worked with post gem, which was a section of on post. And I love I only learned this recently post gem stands for Global electronic messaging. That's how far back we go and pre email all those are pre texts or pre SMS or whatever, maybe. Can you describe to me what post jam was but also of key interest, the set up between them and Ireland's first internet service provider at the time? Ireland online?

Michael Kelly  22:31

Okay, yeah, well, it's a whole subject in itself. But to start with, with post, Jim, it was a subsidiary of, of unparsed, actually decentral as a as a separate company, abroad in people with I suppose a certain amount of it or telecommunications, background plus a lot of marketing, because it was essentially what it was set up to do was to try and develop a new market. And if you if you could cast your mind back to the very late 80s 1989. This is pre mail, or builds went in the post. And that's where on POS came in. But thanks to the foresight of of a couple of very, very clever people, even then they realize that hard copy as it used to be referred to probably wasn't going to be there forever. And the on POS needed to start thinking about the future. And it was it was certainly the first, I think, a pulse office in the whole world that started to think that way. It was very, very pioneering. But when we got the was all very well saying it was electronic mail was very much in its infancy. I think I had you know, I had used it in my previous IT career to probably, you know, communicate with a few other people and maybe with some some vendors, but it wasn't in general use. And it was also quite slow. You know, you didn't get an email instantly wait like we did today. But once post jam got up and running. Initially, I think our first service was people could send in communications like bills or RS circulars electronically. And then we would actually print them and put them into a letter. So it was electronic to hard copy. But it's amazing. That sounds ridiculous now, but it actually got people into the idea. The other thing that we introduced was Electronic Data Interchange, which was electronic to electronic. And basically that was sending purchase orders and invoices from one company to another, completely electronically. And using a set of standards that worked very well. But as become the precursor of what we know now, I would say even pretty much like if you were to go on the various well known websites and order books or whatever, whatever it is definitely the precursor of that. And then also the third service that we we introduced and was pretty much the precursor of electronic mail today. At the only difference was that it was it was a connection rather than to the intranet. It was connected to an A network of other nodes around the country around the world. And only people who were subscribed to those those services or don't note could send and receive emails, it wasn't completely open system like we have today, where you can send an email to someone you don't know or you've never you've never met. So with those three services in mind, they began to pick up traction, and people began saying, okay, and actually, we had a lot of visits from other post offices interested in what we were doing, and then to underlie that. And I suppose maybe this is something that I brought to it, I realized that in order to really make this work, we needed to have our own network at that time, due to legislation and licensing. And so really, the underlying telecommunications had to be telecom era, which as we know, was the monopoly back then. So we got the first value added services license, a value added services license allowed you to offer value ads over a telecom service. But then we went, we said, right, we will build our own data network, our own packet switched network. And that actually became the precursor to our our cooperation with Arvind online. So to answer your second question, Arland online, or IOL, was becoming very, very successful in the in the marketplace, there was a real appetite there for communications. And I think also as well, because we have a certain amount of time, we had a certain amount of insight into the demographics, Ireland was becoming more open. People were emigration, there was a lot of immigration, people needed to communicate. And, you know, instant communication was was what it was all about. And bear in mind, mobile phones were still expensive to you know, to ring somebody for five minutes, or even landlines for five minutes. Whereas electronic mail was free. And you could send as much as you wanted, you could set it say as much as you wanted to mommy or daddy, there were there were no limitations.

Dusty Rhodes  27:10

It was a huge time of change around then, and very exciting. And it was kind of like Ireland was dragging itself out of the darkness of the 70s and the 80s. And all of a sudden, I mean, we we were winning Eurovision, every year, that's what I remember the 90s. All right, and it meant we could do anything. And then we have the football. And we were actually at a World Cup, we could do anything. And then you start talking about this electronic mail, email and mobile phones were becoming more common, as you say. They were very expensive. It was an amazing time. And then Ireland online. I was working with to FM at the time. And I remember, you know, kind of because we were in the younger end of RT, it was myself and Barry Lange. were kind of interested in this internet thing and what it was all about. And then we started incorporating it as part of our programs. And then everybody then wanted an email address. And of course, we were every day on on the air. And the story goes is that Bertie Ahern, who was T shock at the time was listening. And he went, What's that? I want one. And so that's that's it from my point of view in that I thought those kinds of Eddie, very early days of the Internet were quite heavy. What how was it from your point of view introduced the internet to the great Irish public.

Michael Kelly  28:23

It was really, really exciting times. I mean, very exhausting, very, very long days. But we were we really were making it up as we went along. You know, the pioneers and IOL Colin Greeley. And Barry Flanagan like we're real flagbearers for the whole thing. But what I suppose what we wanted to do was try and let them get on with what they were good at doing. And I think what post Shem brought to it was we were pretty good. We mastered the art of infrastructure, and also how we could we could post modems and so on out these days, you'd probably say, well, modem, why would you use a modem. But, you know, back then, the Internet was about getting a CD on that stuck to a magazine, and you'd stuck that CD into your PC. And it gave you a certain amount of software, which allows you to control a modem, which you have to use the home telephone line. And we it was it was quite slow, but it worked. In order to make IOL work and make it successful. We needed to have modems all over the country, because what we discovered very, very quickly was that someone in in Cork or Limerick, probably some of our listeners would say for obvious reasons, they wouldn't be prepared to die of Dublin. They wanted a local number. They wanted a local cork number or Limerick number and so on and so forth. And I think at the end, we ended up with 26 points of protests around the country in order to take in those calls. And we also had to build a fairly substantial backbone, network to form through all of that internet traffic, I think what made it even more exciting was that we just could not have anticipated the demand. As fast as we could put in infrastructure, it was gobbled up. So much so that I was dealing with. And at that time, most of them would have been Silicon Valley based companies that provided the equipment, they wanted to send us the very latest equipment, we'd be the first in Europe to use it, they might only have one or two been used maybe buy AT and T or America Online in the States, they could see that something was really, really happening here. And it was it was growing really, really quickly. And the other thing that we were able to do was because nobody had any real experience in this, we were recruiting from the universe, universities, just graduate engineers, guys that we guys and girls that we taught, you know, which would really enjoy this. And it just threw them in at the coalface and learn what needed to be done. And I can't say there was a plan, the plan probably changed every week. But it worked. And it it's it's one of the things that I'm proud of Southern way all career, I have to say. Because the internet now we moved from a situation where I think when when we went on post Bosch, Ireland online with post Sham, I think we was about 14,000 subscribers. Now there were a couple of other voted, but there was probably probably 25,000 Internet subscribers in the whole country. And that was in 9697. Now everybody use that ubiquitous, you couldn't do without it?

Dusty Rhodes  31:37

Do you have a particular story you'd like to share from that time?

Michael Kelly  31:42

Well, I do, I don't know whether it's a good story or not. But one of our struggles in Ireland online particularly was the connections to the internet. These days, people don't need to really understand how it all works. But back then we needed a connection to the outside world. And bandwidth are the pipes from Dublin to the rest of the world. Were extremely expensive. One stage I we had a 1.5 megabit connection to the outside world. Nowadays, people have, you know, I think over a gigabit into my home, just my house. So we were doing everything that we possibly called we were trading bandwidth with various providers, talking like in lots of money, like it was getting into the millions of pounds at the time. So one of the things that we did do was we we did a deal with a satellite company. Now everybody talks about satellite and you know, watch what Elon Musk is doing. And there's lots of satellite companies. But we had a headquarters on them Earth's for terrorists. At that time, we got special permission. And we did lots of licenses, because nobody was doing it to put this great big, huge satellite dish up on the roof. So Barry Flanagan and I got it working. And we said that we will look let's let's try it, you know before and we'll have it to ourselves, right? See what we can do. So Barry and I were up on the roof. This is a six story roof, where we probably really shouldn't debate to be perfectly honest. But we we plugged in a laptop into the back of it, just to see how it will perform. And it was it was going great and we tried different things. And then I pitched my arm and the laptop fell off the wall or word straight down onto the ground. It's fair to say it didn't work after that. Luckily, it didn't land on everybody. I

Dusty Rhodes  33:37

could just i All the picture in my head is just a pair of you looking. And then there's pure silence. jaws dropped pure silence. It

Michael Kelly  33:45

was it was the longest, probably 15 seconds of my life. Because a bit like Icarus, when things are going badly, time just seems to slow down. A

Dusty Rhodes  33:56

lot of what you're talking about Michael is you're talking about introducing the Internet to Ireland and satellite connections and a one five point 1.5 meg for the entire country. It's ridiculous when you think about it. Now. Another thing that we talk about all the time here, and we're very blase about it because we're one of the huge biggest centers in the world for it with Facebook and Google, Microsoft and data centers. They're everywhere. It takes 20% of the power of the country in Ireland goes just on data centers. You are the man who installed Ireland's first ever data center Tell Me More

Michael Kelly  34:32

probably some politicians would prefer that I had not done this but it yes it is true. I joined I became part of East that telecom as supposed to finish the the post Jeremiah Well story was sold to Dennis O'Brien, you may have heard of. So I moved into to ESA telecom and you know, there was there was a bit of a change around responsibilities and so on. So Oh, Dennis, as, as he did said, Well, look, are you looking for a challenge? Foolishly? I said, Yes. And he said, Well, look, one of the things that we need to do this this thing, data data centers, that's that's going to, that's going to be the next big thing. And I'll be honest, which I was skeptical myself, or really did what he said, but think your demand for this because, you know, it's datacenters are what's going to drive the internet. And I think up to that point, I certainly thought I didn't have as foresight sure, is foresight was that the internet needs lots of power and lots of space disk space, and it needs performance, these availability. And that that basically is a one line summary description of a data center. So he said, right, got to do this. So he said, I think we could get a building out and citywest, which was only really been built at that stage. Now, of course, a very mature business. But back then, it was it was nothing and it had no fiber or any of the telecommunications into its, which was what we really needed. So I think he said this to me at the end of November. And he said, I needed up and running at seven months. Now, I hadn't even seen the building when he said this, and the building was only half built. What that was the challenge. And it was, it probably was the most stressful period of my career. When we did it, by the end of June, with a very, very small team. We built a data center, it was the first of its kind, I did go and have a look at a couple of data centers in the States. I think we brought back some good ideas about it, particularly around availability, we hadn't been thinking about that. So generators was something that we spent a lot of money on. And we also had to fight tooth and nail to get telecommunications in from the various providers so that we could connect these these data service servers. And we got it up and running. I think we had our first customers running probably the month before, and probably about 18 months later was false.

Dusty Rhodes  37:07

Michael, I could chat to you all day. So just let me wrap up one or two little questions about new technologies. Because like everything you're saying, you you literally have been at the development end of everything right throughout your career. So AI is continuous learning. This is what I want to ask you about. I mean, it's not vital for engineers working telecommunications that don't mean is that something nicer is CPD something you should do? Is it vital for working in telecommunications in areas like that.

Michael Kelly  37:37

It's absolutely vital. I think when I started out my career, there was probably a very much of an emphasis on third level education. And it had, it definitely had its at its place. But however you were trained, or whatever you were qualified, and it was seen as a means to promotion. I think organizations are definitely a lot flatter now. And if you want to be really valuable to an organization may be having that master's degree or even a primary degree is not going to help you it gives you some of the tools to learn and maybe to be curious. But I think you know, CPD, getting short courses and technology is really, really valuable, you become much more valuable to an organization.

Dusty Rhodes  38:20

So it's going to ask you was how do you do? I mean, how have you done that CPD just to keep up to because you do you're right, you get your degree and you get on the first rung of the ladder. And that's it, your degree is worthless. After that. You need to keep educating yourself keeping up to speed How did you do at work in such a high level?

Michael Kelly  38:36

Well, the 1990s that is certainly the late 80s 19 1990s, even though we had the internet, and it was still an emphasis on books, and you would you would buy books, and you would learn that way. And you would experiment. Then something changed in the in the vendors, the people who sold the routers or the routers and all this telecommunications equipment, they realized that it was changing very, very quickly. And some of those got into training themselves. Good examples would be Cisco, with their certification schemes, relatively short courses, but they weren't internationally recognized. And I saw people so well, not only would it make them better at their job, but it made them internationally market. So if you know one of our young engineers, and suddenly did decide to go off to Australia or America, they had a recognised qualification. But I think that that's probably brought us up to maybe the 2010s. Now, I think it's it's almost going back. I think you do need the hard qualifications, like the masters and the bachelor's degrees. And so with the speed of change is so much that by the time of course comes out and more importantly is recognized. It's nearly out of date. So you've got to become a sponge. So to answer your question, I think I went from a very much a kind of a rigid book learning type of individual because that's what I was. That's what I was taught to maybe true the experience and so on that you've got to have your antenna, working all the time, operate as a sponge, soak up as much information as you possibly can. Some of its useful, some of it might actually lead you in a different direction that maybe didn't even know was was there. But that curiosity is something that that makes you valuable to the organization. And the more valuable to an organization, the more successful you will become.

Dusty Rhodes  40:32

Let me ask you a humdinger of an awful question. Just to wrap it up. Right. Right, because we haven't mentioned AI. I'm just interested for you who has been so successful, seeing things with potential and then seeing how they could work in the future and then successfully getting them there. You're looking at AI for the last year, we do you think AI will have us in 10 years time,

Michael Kelly  41:00

I think in 10 years time, I think it'll be slower than we thought everybody likes thick. Obviously, there's a lot of hype, I see it first and foremost, in the next 10 years as maybe an eight in your ear, whatever it is that you're doing. Say it could be anything from working in a contact center, it could be a programmer trying to write a difficult bit of code, almost that kind of help or coach in the ear. I see it as specially in them hectic areas. I don't know much about us, if anything at all. But a good example to me would be something like air traffic control, right? If an air traffic controller were to miss something, there's enough technology out there that you know, by tracking people's eyes on the screen, that the AI or some system feeding AI could say, I think he's missed that vital piece of information. Or let I think I should whisper this piece of information in her ear. It sounds incredible Bush a lot of the things that AI can already do were incredible, even five years ago. So I on the one hand, I think it will be slower. But I really do see it as as eight, I do think it will be transformative as well in certain industries. I think the first area where it could really transform is in contact centers, because it AI should be able to deal with different accents, and should be able to deal with different languages, it will be building up databases and other types of bases that are going to apply. And with just the laws of physics, you're able to apply that at the speed of light, either to a screen or you know, to generate something that you call out in the air of somebody, it to me it it has to be transformative. I don't think it's something to be frightened of. That's something that really annoys me. It's like all technology. But I think back probably 200 years ago, when the steam engine was invented, there were probably people given out about that as well. And it wasn't the be all and end all the steam engine transformed into something else. And AI and time will transform into something again,

Dusty Rhodes  43:18

I often see AI as being like the early days of the Internet, which we both experienced in the in the 80s and 90s. And it was the wild west of the Internet back then. Whereas I think we're seeing the A it's the wild west of AI right now. Michael, unfortunately, we've run out of time if you're listening and you'd like to find out more about Michael or some of the topics that we talked about today. There's some notes and link details in the description area of the podcast but for now, Michael Kelly, head of operations of the Emergency Call Answering Service at BT Ireland. Thank you so much for joining us today.

Michael Kelly  43:49

Not at all Dusty, it's been an absolute pleasure.

Dusty Rhodes  43:53

Do remember for advanced episodes of our Engineers Ireland podcast, more information on engineering across Ireland or career development opportunities, there are libraries of information on our website at Also do share a podcast with a friend in the business just tell them search for Engineers Ireland in their podcast player. The podcast is produced by for Engineers Ireland. Until next time from myself Dusty Rhodes. Thank you so much for listening. Take care