Podcast: Automating a city with Copenhagen's Mads Andersen

Seth Adler

In this episode of The AI Network Podcast, Mads Andersen takes us behind the scenes of the City of Copenhagen's automation journey 

copenhagen hero
Photo by Stig Ottesen on Unsplash

Mads Andersen joins us and takes us through the fact that folks see Denmark as ahead in terms of digitalization. However as the head of RPA for the City of Copenhagen he knows that there are silos in the public sectors legacy systems.

Andersen's job is to ensure that the data is found in each silo and combined to expedite horizontal integration as well as engage RPA as the next tool in the box to be more digitized than today. To become more efficient and create more value out of the work which the City of Copenhagen does.

When the city began the automation journey at the end of 2015, they didn’t have anything in place, they started from scratch. Andersen understands the benefit of investing in RPA. He says that he and his team can fail fast and of course succeed quickly.

Listen now: 

Want more? Become a member of the AI & Intelligent Automation Network and we'll keep you up to date with all the latest podcast episodes, delivered fresh to your inbox each week. Sign up today.

Transcript (unedited)

Interviewer: Seth Adler
Guest: Mads Andersen

Interviewer:  Mads Andersen.  And when I introduced myself, you introduced yourself as from the City of Copenhagen and I misunderstood that you mean this is just where you are from and you know obviously where you live but you work for the City of Copenhagen.      

Guest:  I doesn’t live in City of Copenhagen, I live right next door, but I work at City of Copenhagen where I am head of part of the IT department where we try to implement robotic process automation.

Interviewer:  When we first were talking, you said that Fakse expect us to be ahead digitally, what do you mean by that?

Guest:  I mean what I meant when we talked this morning was that in Denmark we tend to say that we are very digitalized in our for example public sector which actually I think is true but well we are not.  We still have different silos within our legacy systems and that is where my job is to kind of combine these things so we can move data from silo to another without… sometimes we develop web services so we can do horizontal integration, sometimes we have the integration platforms and we have spent quite some money on that, some of it was very good, some of it wasn’t and the RPA is just the next tool in the box to compliment these our strive to be more digitalized than we are today.  We are pushing on and on and on to create both efficiency and better quality in the work we do.

Interviewer:  Okay, and so we are at the Nordic RPA and AI Summit and you have this remit to do this thing, how long have you been in the position that you are in?

Guest:  I have been in this position for a year and a half now.

Interviewer:  Alright, which makes you realize one of the old guard, right, I have been talking to it is about 150 people here that is you have been in it a long time…

Guest:  I am senior.

Interviewer:  Yeah.  So what did you walk into a year and a half ago and give us some of the ten poles along the way to your progression?

Guest:  What I walked into a year and a half ago was that we didn’t have anything, we were starting to talk about that automation could be one of the tools to meet our goals, to spend less money on administration and more on education, child care, and to take care of our elderly, so that is because I am in public sector so the less we can spend on administration the better and RPA is one of the tools.  So we started out from scratch.  We didn’t have anything.  We have tried to web services of course and we have a lot of them and a lot of them are good and we can’t live without them but then we had to try to move on without adding the normal cost as web services usually tend to cost.

Interviewer:  So when did you identify bridging the silos as the thing, how and why?

Guest:  I don’t think it is THE thing, it is one of the things to do…

Interviewer: Fair enough.

Guest:  … so that is important to me, I mean we… I co-exist with a lot of other digital things so this is just one of them and one of the what will I say, one of the good things about RPA is that it is fairly cheap to develop so you can actually fail fast and you can actually do, you can automate a process that hasn’t, it doesn’t have to run more than six months and it is still a positive case so you can actually do something and just create and throw away.  And that was the thing that we looked upon on these things, because we tend to get new legacy systems now and then and the next couple of years we are looking into a lot of changes within our legacy systems and starting to do web services now, that is not the best part, the best case, but using RPA until we get a better legacy system, that is a good case and that is why we started this journey.

Interviewer:  Okay.  So little bit less expensive, you can implement and have some quick return.  As far as you said bridging the silos is A thing, totally understood but in that what did you see, how did you come upon that as something that you knew that RPA would work with, how did you come upon it and where did you start?

Guest: Before I started at City of Copenhagen, I worked in a BPO firm where we actually did the payroll for the City of Copenhagen, so and that is seven, eight years ago and we back then we also experimented and actually automated quite a few processes within the payroll department.  So that was what I started saying, okay, so if you are keen on doing automation now, let’s do it, and let’s start with some of the processes I knew that we six seven years ago actually managed to automate.  So yeah, so it was obvious to start within our own share service center and then move along from there.

Interviewer:  Okay.

Guest:  Start taking your own medicine [07:59 ]

Interviewer:  How were you able to implement RPA in bridging those silos as far as specifics?

Guest:  We started out deciding on the first one we did was actually it ended up being way more complex than we thought about but so learning point there but…

Interviewer:  Which one, what are we talking about?

Guest:  … we are talking about when someone leaves, when someone retires, they need to get a letter saying that you have worked here and here from this date to this date and thank you, thank you and then we have to file it in our filing system and we have to create a standard letter and sending it to the now former employee.  So…

Interviewer:  So in “this is a simple thing to do.”

Guest:  This is a simple thing to do, it is rule based and we have all the data we need, we just have to find them. 

Interviewer:  Why was it complex then?

Guest:  It was complex because this small process what was it, five six different legacy systems that wasn’t combined, it was all digitalized of course, no paper involved but it was five six different legacy systems that the payroll consultant have to type in some information to extract something there and file it in a different system and then taking it out, saving it on their own desktop then send it digital to the former employee, so that was many steps but when you talk about the process, it is easy but when you tend to scratch it, it shows that we actually made it quite difficult. 

Interviewer:  Yeah indeed and when you listen to any session, the first thing or atleast at some point, the presenter will probably say what you need is clean data and as long as you have clean data, you are gonna be fine and what you are saying is, yeah sure that is fine, if you have clean data, where is the data, right?

Guest:  Yeah.  It is easy to win the sprint if you are the faster one…

Interviewer:  Yeah sure.

Guest:  So everyone are talking about the golden path or the sunshine story but there is so few of them when you hate real life, you can always create a POC where you can do it.

Interviewer:  Right.

Guest:  But when you hit production, it is something else, and that is fairly okay.  I mean I am trying to sell this and I can also do the sunshine story but we fairly quick have to come to reality and make sure that it is not all shiny and we need to set aside more resources to maintain these processes and getting them to production, making them robust.  I am not saying that it is a negative business case, it is not, but you tend to do it too positive and I am not afraid to say that within my own department because I need to look at these people six months from now.  If I just sung a song, sang a song, which wasn’t true, well it’s my colleagues.

Interviewer:  That’s it, that’s it.  And so it is still maybe less expensive, it still is maybe quicker but it isn’t necessarily perfect and it still takes a lot of work and you talked about maintenance, you talked about kind of upkeep as a true piece to pay attention to yeah.

Guest:  There is a lot.  When you are done automating this process, you have to keep looking at it because it is not if, or when, or it is a matter of when you have to change it.

Interviewer:  Right.

Guest:  Sometimes it is the gooey of one of the legacy systems otherwise it is legislation, it is process, you want to do it a different way, or you find out or there is a million thing that changes and the human workforce is very flexible and they tend to just do things tiny bit different sometimes and they don’t even notice the robot will because it doesn’t work kind alike, it works exactly what it has been told.

Interviewer:  There we go.

Guest:  You can say robots are very, very stupid but they just keep going until they fail, but sometimes the robots don’t make mistakes, the robots fail because something in its environment did change, so you have to keep looking at them.

Interviewer:  Absolutely.  You gave us the example of changing that letter using RPA with a letter, give us another example because again sending letter, simple, small, not so much, give us another example of where you have put RPA in that has been a win?

Guest:  We have concentrated, we have done quite a few things around letters and archiving stuff and also and it sounds stupid printing cases, I started out saying that we are a highly digitalized society but sometimes our tasks need to be printed and we found out that some business department actually spend two, three, five hours a day, one person spends two three hours a day printing out from an outlook inbox and then pushing these tasks around and then we set the robot to just read all these mails, creating one combined PDF and then they could just on one click print it and then they could push the tasks around.  That is the [opposite 14:25 ] so we can also use RPA to go from digital to analog tasks but so we didn’t change the process, we just helped them.     

Interviewer:  Interesting.

Guest:  So that is the different one, which is not a common one but it was okay, it was easy to do and business side was happy, we helped them.

Interviewer:  So the business side was happy, you helped them, that gets into what does this person do with his or her day once we have kind of made that day better with automation.  So can you speak to that the person at that desk that was busy printing for three to five hours, whatever, whatever, I don’t know how detailed and how specific you can get with one person but how has that department or maybe even person improved what they do for the city?

Guest:  Most of the places where we have deployed robots or digital colleagues as I call them.

Interviewer:  Okay.

Guest:  They didn’t have resources to do all the tasks they had to do so they worked overtime or they actually something they didn’t do, now they can manage to do with less hours and do a better quality of work, so at the moment we haven’t looked in, we haven’t seen okay we have fired 15 people or 200 people, I think in not so distant future, we will be seeing that we are cutting down jobs, but most of it I think that we are actually as a shared service center or somewhere else in the City of Copenhagen we actually just do more with the same amount of people.  We free resources to do more complex things because I am not saying that this is public sector, a big McKenzie report says that in Denmark, 27% of public tasks known today can or will be automated 27%, some other industries are looking into 80%, we are not doing that, so and demand is still increasing on, let’s move our spending from cold hands we call them, administrative jobs over to nursery, nursing schools and all that so we are not in the same business as private sector because we don’t have any shareholders, we have voters and we have politicians and they rather spend one million Danish Kroners on better education than one million on us being ten people more doing letters of resignation.

Interviewer:  Sure certainly, however, you did mention what you do the same way the one thing that you attack that the private sector attacks is productivity, right so you have a certain number of people and you just explained how you are increasing the productivity of the folks that you do have and this is something that voters and shareholders both care about right so speak to them.  As far as an increase in productivity, one example is printing, what’s another example where you have increased productivity through the implementation of RPA?

Guest:  Another one was in our finance department where we used to robot to go in and get out information so the controller could do their job.  Before the robot they had to find all these data themselves so the robot doesn’t do an exact job, it just prepares.  The robot finds all the ingredients so the chef can make the perfect dish.  They only have a couple of days, five six days to do this job and before the robot, they tend to do, you spend the first two days collecting data, now the data is served to them the morning of the first day, so they actually got two more days to ensure the right quality of their job or their work instead of only having three so they now get… they deliver better quality because they have more time without changing any process, just being able to get your ingredients or your numbers faster so that is quality to them. 

Interviewer:  Better quality for you, better quality for voters as we mentioned right…

Guest:  Exactly.

Interviewer:  …and so increase quality, increase productivity, what am I missing here, before we get to reducing staff, aren’t these two outcomes exactly what we are going for?

Guest:  We are exactly focusing on, we don’t hide that this is also about cost reduction, but we are also looking to better quality of work because you don’t do repetitive tasks, you get more of your time spending on what would you say the complex jobs where you have to think and not just do so we take out the robot part of a normal day’s job and then you have to do the difficult things which are also the more exciting things to do. 

Interviewer:  Which by the way improves that worker’s life.

Guest:  Yeah, and it also tends to be to create a better life balance because you don’t have these peaks, you can actually use robots to remove the peaks within your work week or work month…

Interviewer:  Or the valleys, right?

Guest:  Yeah.

Interviewer:  Alright.  So you said that you maybe you don’t live in Copenhagen, you live right next door, you said that you came in from… you worked at a BPO, but where did you, where were you born, where did you grow up?

Guest:  I grew up in the City of Copenhagen.

Interviewer:  You did?

Guest:  Yeah.

Interviewer:  Alright.  So you know this city?

Guest:  I do.

Interviewer:  Alright, what was the city like when you were a kid?  Describe that for us?

Guest:  When I was a kid I wasn’t thinking about how it was run, but… 

Interviewer:  Understood.

Guest:  …I was just, I mean, the City of Copenhagen is growing, is growing rapidly, of course compared to New York City or any other major city, we are still small, we are 6000 people but we are looking into a 100,000 within seven years, we are 100,000 more which percentage wise is a big…

Interviewer:  Yeah, gargantuan.

Guest:  …than we are talking about so we need to but just create more administration so if we can tend to use robots for example, to not spending more on administration but keeping up with the same number of employees but just the work is just increasing because we have improved our efficiency with the robots.

Interviewer:   There we go, we are achieving process excellence through increased productivity, increased quality of work so that we can deal with this influx of people, why is the growth so explosive, to what do we attribute that?

Guest:  I think there is a movement, you can call it that people are moving from smaller cities towards the big cities in Denmark at the moment so…

Interviewer:  Okay and so for those that have not visited Copenhagen…

Guest:  They should.

Interviewer:  Alright, and we have got a I guess we are moving on to the next session here at the event but they should why, what is… explain, you are a guy that works for the city so explain it to some New Yorker or some yeah…

Guest:  I mean I love this city, but to me this is, it is an open city, it is a happy city and it is not too big but it has got everything, I mean I just love the City of Copenhagen and I am proud of being part of it so yeah.

Interviewer:  Yeah.  So you… where did you go to school, where did you go to university?

Guest:  I actually, outside the Copenhagen.

Interviewer:  Alright.

Guest:  I was sleeping fairly much in high school so I didn’t have the… so I had to choose a different university.

Interviewer:  And then what did you graduate with, what kind of degree did you get?

Guest:  I have a Masters in Public Administration.

 Interviewer:  Well then this makes sense and then what did you use that in first, what was your first job out of…?

Guest:  I actually never worked within administration.  I started out working with IT even though I haven’t got any signs of IT education, so processes… so as I see IT in process what we say process…

Interviewer:  Process excellence.

Guest:  …yeah process excellence.

Interviewer:    Yeah.  So that’s the language that you speak,  that’s how your mind works.  So to do what you are doing now makes all the sense in the world.

Guest:  To me it does.

Interviewer:  Yeah.

Guest:  My kids don’t understand it.

Interviewer:  Well, how old are your kids?

Guest:  They are seven and ten.

Interviewer:  Oh, they will.  Right.

Guest:  They will.  They are proud saying that their dad works with robots but they actually think it is the physical ones.

Interviewer:  Of course, of course.  Now you did say at the beginning and I want to make sure to clarify, this is just a means to an end RPA, what are you getting at there when you say, we are just using this for the time being?

Guest:  I mean I think within a couple of years RPA would be a common tool in the tool box, but it is not the only one, there will be, we will be more RPA processes, will today be they are fed with instructed data.  Pretty soon we will run out of processes having structure data and then we are starting with machine learning and cognitive agents to feed the robots kind of making order in the data chaos so that is where we are starting to use the next level of technology or so that is going to be a very, very exciting journey. 

Interviewer:  Indeed, but first things first.  Yes.

Guest:  First things first, you have to crawl before you walk and start with RPA and not starting with implementing a cognitive agent because you don’t know what you can use it for.

Interviewer:  There you go.

Guest:  You don’t have the means so starting small.

Interviewer:  This is Cogen advice.  I am going to ask you the three final questions, I will tell what they are and then I will ask you them in order.  What has most surprised you at work and that’s throughout your career, what has most surprised you in life and then on the sound track of your life, one track, one song, that’s gotta be on there, but first things first, what has most surprised you at work along the way?

Guest:  That’s a hard question, what surprised me the most?

Interviewer:  So people often say how quickly things change, people often say how slowly things change, people often say how things don’t change, people talk about the people that they work with, people that you know…

Guest:  That kind of gave me an idea I mean what do surprises me is how we actually managed to change in so different speeds because some people are actually sitting there and the neighbor are just jumping on the train and that is not all positive because we tend to go in different directions because we moved in so different speed but to me I think I am very positive about the City of Copenhagen being a public sector that we are actually openly speaking about us using robots.

Interviewer:  Yeah.

Guest:  That was not, I did not believe that I could manage that with it five years ago, we have fairly strong unions but they are also talking about robots because they know they are coming so they might as well work with them instead of against them.

Interviewer:  So this communication with the key stakeholders, no matter who they are, for you it’s unions…

Guest:  That’s one of the stakeholders, there are many stakeholders but if you don’t communicate, you won’t succeed within your RPA journey?

Interviewer:  What advice can you give folks as far that first step in making sure that folks are in board with you as you make your way into this?  What did you do as far as communication to make sure everybody was there with you?

Guest:  We were very open on why we did it and where we did it and when we did it.

Interviewer:  Just straightforward here is what’s happening, this is what we are doing.

Guest:  Yep, yep.

Interviewer:  Okay.  Fantastic, thank you. What’s more surprised you in life?

Guest:   That’s the biggest question anyone have ever asked me.  I don’t think I get that surprised, I mean I have just, I am fairly positive person, so…

Interviewer:  Just get up, go to work, do the thing, right?

Guest:  Yeah.  It surprises me how much time you spend at work.

Interviewer:  Yes, that’s fair, that’s totally fair.  On the soundtrack of your life, one track, one song that has got to be on there, so doesn’t have to be the perfect song, it doesn’t have to be your favorite song, but on the soundtrack it is definitely a song that is certainly on there.

Guest:  So one song that is impossible but I actually a couple of weeks ago, I was talking about what was the five, name five best albums…

Interviewer:  Fair enough.

Guest:  …and I mean for me one of them are Radioheads Ok Computer, I don’t know why but it is just the sound of my youth…

Interviewer:  Yeah sure.

Guest:  …which just says that I am pretty old.

Interviewer:    Yeah, of course it does, it also says we are probably about the same age.  Yeah.

Guest:  Yeah. So that could be one of the few like major things and of course, of course, if one song, it must be as most lightening spirit.

Interviewer:   It’s Nirvana.

Guest:  I mean that’s Nirvana because that was where my life kicked on, I mean I was in my youth and I was heading forward pretty fast, and I was actually living in Seattle in 1994.

Interviewer:  What were you doing at that time?

Guest:  I was studying.

Interviewer:  Okay.

Guest:  So I landed there four months too late because Kurt Cobain actually managed to die before I arrived.

Interviewer:  But this is still, I mean as far as Grunge…

Guest:  This is probably the song, there are lots…

Interviewer:  Sure, but don’t …

Guest:  …so don’t get me started.

Interviewer:  …I can’t, I won’t but we will keep that song absolutely but four months after he died, the Grunge scene was still huge at that point, so don’t sell yourself…

Guest:  It was exploding.  No, no.

Interviewer:  … sure and Chris Cornell just died the other day…

Guest:  Yeah, tragic.

Interviewer:  …so that is RIP, but fair enough I guess we have got a lot to talk about as far as music once I turn these microphones off, how about that?

Guest:  Yeah.

Interviewer:  Mads, it has been a pleasure.

Guest:  Thank you.

Interviewer:  And there you have Mads Andersen pointing out that first foray into automation turned out to be a little bit more complex than initially anticipated.  This small process involves five or six different legacy systems, the human kind of just connects them and changes how he or she does it.  The robot notices a think like that.  Thanks to Mads, thanks to you, stay tuned.