Volvo Car is pioneering financial robotization
Anton Edlund, Volvo Car Group Automation Manager, Intelligent Process Automation
We quickly think robotic process automation (RPA) is food for high-tech startups and academic wizzkids. And yet it is the industrial giant Volvo Car Group that is in the front row when it comes to real robotization implementation. ‘As long as something is rule-based and repetitive, you can have it done by a robot,’ says Anton Edlund. He leads the global Intelligent Process Automation team that automates financial processes. During a visit to the Volvo factory in Gent we can put him to the test.
When we talk about Volvo Car Group - owned by the Chinese company Zhejiang Geely - we have to figure out big numbers: sales of around 24 billion euros, 1.3 billion EBIT, 642,000 cars sold, 43,000 employees. They are spread all over the world, and that implies that processes must be streamlined and shared throughout all businesses. This is no different in finance. There is the ‘Digital Finance’ team, about 50 people. This includes the 10-head team ‘Intelligent Process Automation’, which is led by Anton Edlund. In a mixture of intelligence and typical Swedish modesty and human centricity, he explains what he and his people are doing.
First things first: what do you mean by robotization?
Anton Edlund: ‘We are talking about software robotization, also called RPA. With our team we try to automate as many manual tasks as possible within finance, to support our teams around the world. Think of administrative matters, but also more complex matters such as monthly and annual closures, reconciliation, reporting, etc. As long as something meets two conditions, we can robotise it: it must be about rule-based and repetitive tasks. If one of those two things is not met, robotization is still an option, but then a lot of artificial intelligence is needed to make such a process a success. I immediately see this as one of the major challenges in the future of robotisation. But understand me well: we are not there yet.’
Why is Volvo so resolutely pulling the RPA card?
Edlund: ‘In the first place because we are absolutely global. China, Europe, USA: there are no borders in our company. Moreover, we automate processes in order to be able to share as much knowledge as possible across various functions. So it is definitely not just about finance. We design a software robot whose ‘engine’ can also be used in other units, such as HR and supply chain. This ‘knowledge sharing’ is a crucial reason for us to robotize. All our team's efforts are therefore shared openly and transparently throughout the company.‘
That sounds easy. Will this international cooperation also succeed in practice?
Edlund: ‘Transparency is the key concept here. It requires that we are open and on the same line. Feel free to call that a thoroughly Swedish mentality. The major advantage of this transparency is that it will facilitate adaptation to changes. This is no different for robotisation projects. Everyone in the business can follow how our work evolves. Our way of working is based on mutual trust.‘
What does your team automate?
Edlund: ‘Gosh ... ask me earlier what we don't automate. Our team only started at the end of 2017, but in the meantime we have been able to automate a lot of detailed, rule-based and repetitive tasks for many businesses. Our ‘shop’ with robots is actually quite full after such a short time. And thanks to our transparency, anyone in the entire company can look around in that shop and find out if he or she can start something with those automated processes in their own business.‘
It seems like a job to robotize financial processes. How do you get started?
Edlund: ‘We will tackle this in phases. First of all you have to understand the process very well, before you prepare a roadmap for the robotization process. In the first step you also have to dare to honestly ask yourself: is it a good idea to build a robot for this process? To investigate this globally quickly and objectively, we have developed our own survey tool, with qualitative and quantitative parameters. Only if this assessment shows that robotisation is a good plan will we proceed to the next step. Because automating a bad process is of course foolish. When developing a roadmap, we look at how we are going to evolve from ‘as is’ to ‘to be.’’
What is the third step, once the roadmap has been drawn up?
Edlund: ‘We dismantle the process in the smallest possible sub-processes. Automation therefore takes place at the smallest possible level: more at the micro than at the macro level. We have developed our own standardized methodology for robotization itself.‘
As a person, do you not risk losing insight into the process?
Edlund: ‘No. Because we always take into account a human fallback solution during automation: people must be able to intervene at any level of a process if the robot fails. It is therefore crucial to continuously understand what you are automating. That is why we at Volvo attach great importance to the fact that the inquiring business continues to be the owner of the process: ownership remains in the business.‘
What are the direct profits that your RPA projects generate?
Edlund: ‘By automating at the micro level, you design small building blocks. Short sprints, carried out by experienced experts in our team. And those small blocks have the great advantage that they can be reused in other processes, in finance but also elsewhere. The smaller your building blocks, the better you can assemble them afterwards to design a specific robot. At the start of our team in 2017, that was the challenge: building blocks that are reusable. Well, I think we've won that challenge. And our objective is clear today: the more building blocks we build, the larger our library becomes, from which we can draw to build robots.‘
If we look at, say, 2030: where will Volvo stand in terms of robotisation?
Edlund: ‘Honest: that is unpredictable. Robotization is evolving at lightning speed. The only thing we can do is follow these developments with full attention. The system landscape and the applications,…: everything is progressing fast. So my message is: follow up and stay modest.‘
Then there is the human aspect: robots will make all kinds of finance functions superfluous.
Edlund: ‘I don't believe that. More and more data is coming our way. The need for human control will continue to exist. I therefore rather see a shift in the task of the accountant: this will evolve towards more control, somewhere between accounting and controlling. And secondly, I think that new finance functions will be added, at the interface between IT and business.‘
Give an example of that?
Edlund: ‘In the past, digital solutions were typically built by IT people. But its development has become so easy that they can be built in finance itself. So digital and technical functions will come in finance. That will be a nice mix. And in this way, finance itself can also create more value.‘
I hear you say remarkably little about artificial intelligence. Is that conscious?
Edlund: ‘First walk, then run. You must first learn and then do it. Machine learning is only ‘learning’, not ‘doing’, as the term itself indicates. It is still early to put full machine learning in a robot, but in the future we may not need robots anymore because AI not only learns more but also does it itself. In short: we keep a close eye on AI, but first want to see what it will learn.’
(Pictures (c) Jerry De Brie)