Blueprint Software Systems—Interview with Leslie Willcocks
Interviewer: Matthew Agnew, Blue Print Software Systems
Interviewee: Leslie Willcocks, LSE and Knowledge Capital Partners
Date of interview: 23rd January 2023
Matthew Agnew: Hello, everybody! Thank you for joining us today for another addition of Automation Untangled. My name is Matt. I am with Blueprint Software Systems, and today I have the great honour of being with Doctor Leslie Willcocks, of the London School of Economics. Thank you so much for being here today. You have written a lot of books—any favourites?
Leslie Willcocks: I always think the first book is the favourite, and that that's quite interesting, because it goes back to 1987, Computerising Work. In those days we were busy computerising work, and we weren't talking about computerised work, and the issue was everyone didn't know what happened when you brought it into an organisation. It was quite a new phenomenon, so I did several years of project management before writing the book, and I realised that no one really knew the human and organisational dimensions of introducing it. So that was when I got really interested in the fact that this is going to be quite complex, and it was going to go on for a very long time, and it was always going to be relatively difficult. And so it is proven!
Matthew Agnew: Absolutely. And Dr Willcocks is an authority on really everything to do with information systems, and especially newer information systems, as you mentioned, and I think a lot of that has to do now with automation and robotic process automation. So can you tell us a little bit about how you got started in the field of robotic process automation.
Leslie Willcocks: Well, in 2014 I didn't even know what it was. The term was invented in 2011 / 2012 by one of the vendors - Blue Prism. So, they invited me along and said, “Look, this is really interesting stuff, this isn't hyped, this does good stuff. Could you have a look at some of the cases?” So that set us off really, and I realised that there was real ‘meat in the sandwich’, as it were, that this really did, to coin a phrase that I used subsequently, ‘take the robot out of the human’. It automated very simple tasks that humans didn't like doing, and could do it very easily. it was a waste of imagination and creative energy and intelligence for them to be doing this kind of work when they could be doing higher level work. I could see the real value of this. We researched it and found that it was producing a potentially immense value for organisations - if they utilised it properly.
This is one of my one of our great findings of over 45 years working with computers: this stuff is grossly under-utilised. The potential is massive, and then you go into many organisations, and, at best, 30 percent of the potential is being utilised. The result is this: a massive amount of value being left on the table. So, I've sort of been devoted really to understanding the technologies, making sure that there is ‘meat in the sandwich’, as it were, and then finding out how the leaders manage these technologies to get the potential value. I mean, if you look at RPA. At the moment, I'd say that the top 20 percent of users of all of RPA are getting about 90% of the total value. So, as far as there has been value globally from RPA, it's about 20% of the users is getting it because they manage it very well, and there's a lot to that - we've written books on it. But the question then, is, what are the others doing? Well, they're very often blaming the technology!
Matthew Agnew: It's funny you mentioned that because you know you said a couple of words there that really are triggering my thoughts. You said ‘potential’, ‘people that are doing it right’. I think sometimes with RPA. I get this sense that there's often two steps forward, one step back approach where there's a lot of problems people are facing, because they were so excited to do this, with so much potential value, that they jumped into it. So, with this potential value, and with everyone jumping in, what are some common issues that you've seen, or where you think “You probably shouldn't do it that way”?
Leslie Willcocks: Well, the first one is that, when you look at RPA, it's a relative no-brainer, it will give you quick benefit. You bypass the IT department. The users can see the impact quite quickly, and they buy more. The outlays are relatively small, and the piloting, as you know, is relatively short—weeks. You don't need a lot of skill—technical skill—to get the thing up and running. Most of the RPA vendors would tell you that. Then the benefits come through. And then there's the question of what do we do next?
This is where most organisations are. When you look at the surveys on this, you know how many actually scale up beyond one to 50 robots, as they call them, or software licenses. And the answer is not that many. Because they reach a point when they say, “Well, we've got these benefits, but we're now incurring some costs to get those benefits. If we scale up the cost benefit analysis doesn't look great because we're going to have to do a lot of things we so far haven't had to do to get the benefits, and maybe, you know, we haven't got the capability to do this, and maybe there are other things that we should be investing in.”
So, the second stage usually is not entered into, because there's a big barrier. And you really do, at that point, have to think about managing scale. How do you manage scale, and how do you persuade senior management to invest in scale when the benefits are not going to be immediate, but could be very big? If you then have a strategic view of these digital technologies, and it links to digital transformation, these benefits could be massive and fundamental.
So, there's always a barrier there at that point and surprisingly few organisations get over that. I found that organisations are not all the same in the way that they utilise automation and do digital transformation. We'll talk about the present environment in a minute, and what impact that has. But you know, I see organisations in any sector falling into four kinds. There are the leaders—about 20% across sectors—who are very good at automation scale up, make forward investments, tend to be more strategic in their approach and link it with digital transformation technologies—integrate them.
There are followers who are trying to do that. But don't have the management capability, or the size of investment, or get hit by other problems.
Then there are the laggards who are really struggling with their automation and might not even get beyond the scaling up challenge. Then there are - it's an unkind phrase, but one could call them the ‘also rans’. But basically, they don't really have much idea and that might be because they have so many other things going on in the present economic context. You can imagine a lot of organisations don't have automation on their mind. What they have on their mind is: how do we survive from a day-to-day basis? And so far as they use automation, they’re sort of sweating the assets they've got. They’re not looking to invest anymore, because they haven’t got the money, they haven’t got the cash flow, that's what they're struggling with.
So, what we're finding is that we study the leaders and then we study how we can help the laggards, the followers and the also rans, to move from where they are to get more benefits from automation and digital transformation. So, it's quite a complex picture.
Matthew Agnew: I'm interested ... you know you mentioned digital transformation a few times? And for me you know that phenomenon when you hear the same word over and over again, it kind of like if you say buffalo, buffalo, buffalo, buffalo, eventually you can't hear anything, it doesn't sound like anything. I find, for me, I've been in the IT space for a while, and digital transformation is kind of one of those terms for me where so many people have said it for so long. It's this buzz word. For you, what does digital transformation mean?
Leslie Willcocks: Well, it means what most people AREN’T doing!
Matthew Agnew: I love that!
Leslie Willcocks: But they use the phrase to describe what they ARE doing. So, digital transformation is a whole organisation, radical restructuring of people, processes data and technology, to become a digital business, essentially. Most organisations are really having problems with that, because they have what I call ‘silos’. The siloed organisation. You could go into most organisations and say which of the following silos: do you have? Have you got data silos; technology silos; culture silos; strategy silos; managerial mindset silos; process silos, skills and structural silos? And most organisations will say, to varying degrees we’ve got all those. And some will say, well, seriously, these are real barriers to digital technology.
Digital technologies do not like silos. And those are the legacy silos of the organisation because the organisation was set up to run in a different way, and then you bring in technologies which allow you to run in a different way again. But you need to reorganise and restructure in order to allow those technologies to reveal their potential.
So that's what digital transformation is to me—breaking those silos down and utilising the potential digital technology to become digital businesses. But I think a lot of people have redefined it.
So, for example in the recession that we've got at the moment, because and all the signs are that organisations are really cutting down on investments, that they over expanded, and thinking that the technology would take off, because of Covid giving them positive, warm signals about the role of technology in a future economy. I think that they now talk about cost transformation through digital technologies. And that's the way that digital transformation is being reconfigured. The terminology is the same but what it means is how can we use digital transformations to keep our costs under control and achieve economies and process improvement without really doing those fantastic strategic things or using the top leading-edge technologies that people have been trying to sell for that decade.
So that's what where we are at the moment I think, even the go-head companies, the leaders, seem to be coming back into defining digital transformation in practice as cost transformation. How do we transform our cost space? So, there's a role for RPA, and intelligent automation, and AI, but it's much more redirected into narrow goals.
Matthew Agnew: Yeah, and you know, one thing I think of is, so for taking RPA and intelligent animation, we're putting it under this umbrella of digital transformation. One of the first digital transformation, common initiatives, that I remember 10 years ago was this concept of digital signage. So you know universities want to have this ability to quickly have digital signage and things all over the place, and I remember speaking to someone at a university that said: I set it up, I got the programs, all the monitors are up, everything's up, but I don't know what to put on them. I think sometimes the same thing happens with RPA. Where it's: We're good, we have the tools, the developers, we know what to do, but what do we automate?
And so where do we find this link between business and the developers and the IT? Where's that kind of sweet spot where we can make this work right, because we always hear that with RPA: we automated the wrong thing; or we didn't fully understand the process before we automate it. And the big concern about that is, when you do that, you completely blow out these benefits that you're talking about. They're gone because you're going to spend too much time maintaining it and fixing things after. So where is this sweet spot, this link?
Leslie Willcocks: I think it's before the sweet spot. I think it's where you start from. I remember, back in the 1990s there was a big thing about business process reengineering, and you know, radical reform. I think it was overstated in terms of the radical nature of it, but I think that you've got to start with the processes of the organisation. You got to identify what the core processes are, and how you're going to digitise them and then the other processes as well.
So, for me, it begins with process discovery, really. What are the core processes given the sort of business we want, and how do we utilise digital technologies to leverage those processes? And how do we organise our data to run those processes? That, to me, is the fundamental starting point. And how do you do that in an organisation? I don't think you can do it from a bottom-up approach, but I don't think you can do it from a top-down approach—which is where radical, business process reengineering went wrong. I think you have to do it from the middle. Because somehow, it's a shorter communication path, and it links with the bottom-up people who are saying, “Well, we're really having problems here”, and the top-down, people said, “Well, strategically, we need to go here”, so you then say, “Well, what processes do we need in place?”, and you invent some form of governance that allows you to do process discovery in different parts of the business. What are key business processes to deliver on your business imperatives? And then you start organising your data and your technology around that.
What I've seen with robotic process automation is it's been sold as a bottom-up approach, and each time it's a one-off process. You know, “We've automated this and look at the benefits we're getting from this process. So, let's look at another process!” But do we ever look at it from a more holistic perspective? People tend to automate the processes easiest to automate first, but is it more strategic to look at the ones that are going to give you the most value?
Matthew Agnew: Yeah. And I find sometimes there's always this disconnect. I always call myself tech adjacent, where I'm not an IT guy—I understand technology, I can't do a lot of the things, but I always find, when you look at things from a business perspective a lot of times there's a conceptual approach right? You know what we want to do. And then with IT, teams and developers can be very literal and sometimes there can be problems there.
I remember one of my first jobs. I was pretty young at working as a product manager, and for the solution I was offering some suggestions based on the customers, and I had said to the development team, add a button that says, ‘save changes’ or something like that, and that's what I wrote in my notes. They updated the product, and they sent me the spec, and I looked at it, and the button said, ‘save changes or something like that’! It was exactly what I had said. It wasn't my fault; it wasn't his fault. It was that the process wasn't in place to have this governance like you said, and really say, this is how we do it. This is how we take our conceptual idea and give exactly what the developers need to get their job done. And I think that's where there's this big push now for this automation, CoEs—Centres of Excellence.
Leslie Willcocks: Well, we concluded that back in, you know, between 2016 and 2019, and then we saw the Centres of Excellence, sort of moving from just RPA Centres of Excellence, to automation Centres of Excellence, and then them linking up with digital transformation efforts. What we found in the big companies, certainly, was that there was this hole, which was automation initiatives were not connecting up with the team running the digital transformation initiatives. New silos were occurring in these big organisations, so the digital transformation tends to be driven top-down, and the automation tended to be driven from the Centre of Excellence at best, and very often bottom-up. And there was little integration of the automation technologies - there people were trying to get into the what we call intelligent automation —what people like to call AI, but I call cognitive, because I’m not a great believer in the use of the phrase ‘AI’ at the moment. The technology being used in most businesses is not that advanced. At the very best it's what I would call ‘weak, weak AI’ as far as I can see, in terms of what most businesses are using as opposed to what's going on in research.
So yeah, I'm finding that people are going digital but in a relative fragmented way, and that’s a problem. In terms of the sweet spots, first they have to get fundamental about process which is best driven from the middle of the organisation, linking it with business imperatives. Then you organise your data and technology around that. And then you've got a link from the middle of the organisation with the digital transformation efforts in order to start integrating what you're doing with the automation, and what you're doing with the other technologies. Because, as you know, digital technologies really gain their massive value from integrating with other digital technologies. When you start linking blockchain with automation; when you start linking seriously advanced automation technologies with RPA, you suddenly start getting much more value. You start getting what I call a platform—an automation platform—then eventually a digital platform. And it's the platform that enables, not just new business processes, not just new business development, but new businesses can spin off the digital platform, as we've seen with the obvious cases—not least Amazon.
Matthew Agnew: I find that RPA is following the trajectory of a lot of other big IT initiatives. I remember there was this phrase going around with cyber security, where, if it's security, it’s everybody's job, everybody's responsibility, and I remember my thought of that as always, well, if it's everybody's job, it's nobody's job. So, you know, the other companies did have a CSIO, and these security people. I think of RPA going that same route where like, oh, It's everyone's job. It's a part of this job; and then there's somebody in business that's helping; and someone in finance is going to jump in. Even today it's very hard to look at a company and say, who is the number one RPA person, who's in charge of this, and I find we're kind of causing that same problem that we have with other initiatives. Everyone’s job ends up as nobody’s job.
Leslie Willcocks: Well, it's a bit like if you follow soccer, or a game like that, you know it's like saying, well, we all have to defend. But what are the rules of defending? Whom am I marking? Where should I stand? It's a bit like that to say it’s everyone's job, which is why one of our fundamental findings from looking at RPA for the first eight years was that you have to centralise governance more than you think. That's why we keep saying you have to start in the middle. You can't really start bottom-up. You know it's nice to try out a few things, but that's a very low-level fruit that you're picking there and you have to start much higher up in the organisation in terms of governance.
Matthew Agnew: Very good advice. So again, I'm thinking, you know you’re considered an expert in this field, so I imagine you get this question a lot. What should you consider when you're looking at an RPA tool? I'm not asking you which one's best, or which one you recommend. But what is some guidance of when someone's asking what tools should I use for my organisation?
Leslie Willcocks: Well, the first thing is that they are surprising differences between RPA. tools. A lot of people don't realise that, and they might go for the cheapest, because they say, well, it's all crisps—all crisps are the same, so you think all RPA tools are the same. They're not. They're very different. Some actually do give you enterprise RPA, where you connect up with the enterprise systems, and a lot of the problems are taken away from doing that, so you don't need a lot of development. Some are just basically frameworks, which you then have to develop the software for. Some are just basically organised around the desktop and don't really do enterprise automation. And then some are in the cloud—you bring them down, and they're relatively easy to use for certain, limited tasks.
So, the first thing is to ask is - what are you looking for here? If you're if you're just looking for a desktop enabler, then there's a bunch of them. What are you trying to achieve? What is your purpose? That is the first thing.
The second thing is the pricing. What are you actually getting for this very cheap price? Very often the cheapest price is not really the cheapest, because there are so many add ons later, and also, if it's very easy to use, very often there are large scale maintenance costs that come later, because it was far too simple, really, for the complexity that you want to apply it to, eventually.
I would like to reiterate: You have to have governance from the middle of the organisation very early on to drive this through. You've always got to be able to reverse out of your suppliers. I mean, this is just the thing of mine and it's not specific to RPA. I know that it sounds highly cynical, but not everyone stays in business. Tools are not always appropriate all the time for you for what you're trying to do. You might outgrow the tool. So, you've got to have the ability to change suppliers in some way, and you should always build your internal capability. Most of the organisations that I know do use external help, because initially, they don't have the capability. They do use consultants, but they use them in order to learn and build an internal capability which frees you up a little bit from, and it gives you more control over, your automation destiny.
It's also very difficult to deal with vendors if you don't know enough about the technology. Someone in the organisation has to know something about the technology. You don't know what you're buying and you don't know the limitations of it. I mean I buy things that are not particularly good all the time, but I know the limitations of what I'm buying, and so I know what the usefulness is, and I know that, despite this, I'm going to get benefits, so you can always have to bear in mind that these things don't arrive 100 percent perfect or a 100% fit.
Matthew Agnew: It reminds me of that adage I learned when I up being a product manager of: From good, fast, cheap, pick two—you can't get all three. So it sounds like some people are picking the fast and the cheap. You know it’s not going to be good, and then they run into some issues down the road.
What you were just talking about—being able to reverse out. You know that's kind of a blueprint in focus right now, and it is challenging for a lot of people because there are two things that I've noticed with this idea of being able to reverse out of a vendor. It's people just get so stuck, and they put so much effort on these older tools or the original tool and when they look at the effort to move off of it, they just think that's too hard. And the other thing is, people are scared to recommend something else, because a lot of times somebody can get fired for recommending the wrong thing. So, it's a scary thing! What’s your advice for when someone is looking to switch a tool? What are some other kind of added devices? Someone's maybe thinking, okay, I've been with this one for 5 years, I'd like to go to another one.
Leslie Willcocks: Well, I've sat in on those organisation discussions many times, going back a long way. It's always difficult, internally, for the organisation. There's lots of umm-ing and ahh-ing. There are people who like the tool and own the tool. And you say, yeah. but you know we want to do this now, and it doesn't fit, whereas this other one does fit, and was designed to fit. So, we're going to have to move on.
You always have to manage the internal politics of a changeover, and you always have to provide a very strong business case to get that switch over. You really have to have top management on board, because only they can really sign off on something that looks like an additional expense, and you have to show that over two or three years, it's not actually an additional expense, it's a reduction in expense, because otherwise you're going to write an awful lot off to the maintenance budget.
I mean over the years it must make you smile when you hear the word maintenance, because you know as well as I do that it’s just a euphemism for a real dumping ground for lots of failures, mistakes, inefficient systems, people covering their ass—all sorts of things go on under the heading of maintenance. When I go into organisations, and they say, well, what do you recommend? One thing I i say is - can I have a look at the maintenance budget?! And I always find an immense number of sins going on there that you can help them with.
Matthew Agnew: I've seen some. I won't say a name, but I spoke to someone that runs RPA at one of Canada's largest insurance companies. Their processes are very good, but there's one thing that that scared me. Their process is, essentially, if a department wants to automate something, and they go to this RPA team. The RPA team comes in, they find out what's going on, they spec it, they say, okay, it'll save you 3 hours a day, it'll cost us this much to do it, and this will be the ROI, right, and they estimate the maintenance cost. Then they build it, they automate it, they test it, they deploy, it, they high-five each other that it's all done, and then they never think about it ever again. And then, meanwhile, the IT team is dealing with the maintenance and dealing with the not running and having to rebuild things. And I just find you've created this cycle of errors, and I know again with the CEO's people have gone over that. But do you think is that still a common thing, or most businesses kind of past that point?Leslie Willcocks: No, it's, unfortunately, still a thing. Businesses frequently don't pass those points, and very often a whole load of new people come in trying to understand what they've inherited. And so, there's a delay in actually resolving those problems. And if they're very large organisations, very often people will leave again, and so they they're not resolved, they're just kicked into touch as it were. So, organisations are complex places to work, as you know. My sense is, from the way you're describing it, that I probably have studied that company! It's in our database. We've got about 1,100 plus organisations in there, including a bunch from Canada.
You do need to get experts in at that point because it's not going to be resolved internally. There would be too many people who don't quite understand the problem or haven't seen it before in quite the same way as experts have. It might be more easily resolved if you sort of put an expert in and I’m always surprised at what real experts—REAL experts—can do with IT.
I remember I used to, for my sins, be fairly expert in outsourcing contracts, and I thought I was pretty good, and then I came across a guy who could come into an outsourcing contract and reduce its costs, without any noticeable adverse impact, by 25 to 30 percent— in an on-going outsourcing contract—just by looking at the contract, looking what it was actually being delivered, looking at the pricing, how outmoded it had got, whether they were getting paid for things that they weren't actually doing, all these sorts of things. I was absolutely surprised by it. And he, in this particular case, actually saved the organisation $10 million dollars on a 5-year contract—which is quite a lot of money—for his fee outlay of probably $30,000.
Matthew Agnew: There's going to be more need for those kinds of things moving forward. People are looking for more ways to save money. There are all the signs, with recessionary conditions. I think that RPA is going to be a big part of that in two ways. Obviously, people are going to need to automate some tasks. It might be less employees around. And then the other thing is, they're going to need to tighten their budget. I think that once they look at letting a few people go, then they can. They look at each department and say, hey, shave off some money. So, I think these RPA teams have a lot of money and expensive tooling and things like that, so they're going to be looking for ways to save.
Leslie Willcocks: I’d just I like to add that I'm a great believer that we don't exploit our existing technology very well. Recessions/downturns are times when you should look at how you are utilising the technology, and whether you're fully applying it the way that it could be applied to your organisation or a revamped process.
So, there is that. But I have also spent a lot of time with organisations identifying the technology that's not paying off—and never will pay off—but on which you're incurring a large maintenance budget. It's surprising how much time and money that those sorts of technologies are taking up in our modern organisations. Yes, it’s part of the legacy inheritance. If they could get rid of those, then they would tie less people up. Instead, push them into developing better technologies and applying them better.
So, it's a time again, now, present circumstances, to start re-looking at whether the technologies you have got are really doing the job. and if not, why not?
Matthew Agnew: Dr Willcocks, it has been fantastic talking to you. I know when we set up this call, we had this long list of things that we wanted to talk about. I think we got to quite a few of them, but I'm sure we missed some, so, I would love to speak to you again sometime.
Leslie Willcocks: Yeah, let's do that. There is so much more, and it’s also changing very fast these days.
Matthew Agnew: I'm sure we peppered on some. But I truly feel like this was a valuable conversation, and if you're in the RPA space. I really hope you learn something. So why don't you tell people where they could find you, or learn more about you, and tell us about one of your more recent books as well.