“We’re Hiring Robots, or How to Automate Processes in Fintech”
Mariusz Pultyn (CTO, Digital Teammates): First, we’ll say a few words about who we are and what we do.
Przemysław Kamiński (Country Manager, IPF Polska – hapipożyczki / hapiloans brand): Hapiloans brand is a part of the international business IPF Digital. Currently we operate on 9 markets, from Finland to Australia. In Poland we offer our customers quick access to money, to long-term loans. As a fintech business, meaning financial technology, we seek solutions that facilitate optimisation, streamlining, acceleration of processes so that we can use all the new solutions available on the market to function faster and more efficiently, to be more customer-friendly, so that the customer can get the service quickly, simply, easily, effortlessly, without leaving the house.
Mariusz Pultyn: I’ll briefly say what Digital Teammates is. Our mission is to automate boredom at work. We automate boredom with the help of our robots. You can think of us a bit as a temporary work agency. Our robots come to work after they’ve been trained, they are ready and eager to do things which are boring, repetitive and not very fulfilling in businesses. We do that without any upfront costs and, which is probably the most important, we guarantee that our digital teammates come to work ready every day.
Przemysław Kamiński: As our business grew, we were faced with a challenge: how to cope with the increasing number of applications, which our call centre had to handle, and as in any other business we, too, were considering the cost aspect. We asked ourselves the question whether “more” has to necessarily also mean “more expensive.”
Aleksandra Rzeszutko (Process Optimisation Head, IPF Digital): We also wondered how to diversify our employees’ tasks. We have more and more young people in our team. Nowadays clicking isn’t really interesting or appealing. We wondered what to do about it.
Przemysław Kamiński: Additionally, the traffic of incoming applications isn’t regular. There are moments when the applications pile up. So we were presented with the challenge to find a way to relieve the traffic that occurs at the call centre.
Aleksandra Rzeszutko: We could implement overtime, but the work-life balance is crucial to us. We didn’t want for our employees to spend their free time at work, solving our problems.
Przemysław Kamiński: So we thought of robots. But there was a risk that the moment we implement new processes in our business, it could negatively affect the core business.
Mariusz Pultyn: Fortunately, the solution we offer works as I explained. We are a type of a temporary work agency, which means that we solve problems, not add to them. Which is why we take responsibility for the robot’s whole work process – starting from how to recruit it, how to train it for work, and finishing on making it go to work every day – and, sometimes, every night – and do its job in a predictable, ant-like way.
Aleksandra Rzeszutko: We were terrified that it will probably be another IT project. As we know, some projects take a few months, some take even longer. We wondered how to do it in a cool, quick, right way.
Mariusz Pultyn: We usually aim to deliver the first robot within 6 weeks from the start of cooperation. Here, thanks to the fantastic support and receptiveness of the hapiloans community, we could reduce that process to 3,5 weeks.
Przemysław Kamiński: Yes, IT projects, as it happens, can be very expensive and, as Ola said earlier, take a long time. We managed to complete this project in a few weeks, but, still, before starting it we wondered how much it’d cost us.
Mariusz Pultyn: Yes, indeed, usually you need to have a lot of funds before an IT project starts bringing any business benefits. It is an issue and when we were beginning our adventure with robots, we knew that we had to propose a model in which our clients could benefit from the first day from robots working faster and more efficiently. We decided on a subscription model in which the robot receives its pay, a kind of salary or rent, as you prefer, only when it does any work of value. And, just like any other employee at the end of the month when they complete their work, the robot can expect a pay consistent with what it managed to do.
Przemysław Kamiński: From our point of view it’s advantageous insofar as we don’t pay social insurance.
Aleksandra Rzeszutko: To us our customers are the most important. Because we are a financial institution, the safety of our customers and their finances is also crucial. We wondered how to solve that aspect.
Mariusz Pultyn: Indeed, that was also one of our objectives. We wanted to be sure that our clients could trust us. From the very beginning we decided on a model in which we guarantee that our digital teammates will come to work and do their tasks. We give a guarantee that the robots show up to work every day.
Aleksandra Rzeszutko: So we talk about these robots but what do they actually do for us? As I mentioned before, we wanted to choose such topics, such processes or tasks which are tiresome or repetitive. Therefore we decided on the 1-zloty verification transfer process. Until now, every now and then an employee would log onto a relevant program, verify received payments, then follow the next steps in our process. Thanks to the robot, which we hired for this task, we get this information on an ongoing basis, and our employee can immediately proceed to verify the application.
Another such topic was the management of the loan application list. Here the applications are also already automatically verified by the robot and passed to our agents at the call centre. As Przemek already mentioned, thanks to that we send money to our customers faster because the robot works 24 hours a day.
Another area where we implemented robots are all kinds of jobs, that is updating the status of tasks in systems. Sending contracts, invoices. We send repayment reminders thanks to our robots. Another task is generating all kinds of documents our customers need. And in Mexico we decided to use a robot for calculating the credit score. Right now we have 40 robots, our new teammates, who operate on 15 processes.
Przemysław Kamiński: It is worth noting here that we have robots who control other robots on other markets.
Mariusz Pultyn: Yes, indeed, in a moment I’ll probably say a bit more about robots working as controllers for other robots. Now I’d like to explain what happens in those 3,5 weeks, what occurred in hapiloans in those 3,5 weeks.
As you can see, the to-do list to complete before a robot can go to work is quite extensive. I won’t discuss all these tasks because that would take a bit more time than we planned for the whole webinar, but I will focus on some stages of this process.
First of all, in this first task, before you can hire robots, you have to at least realise where these robots can be useful. Experts call this the robotic potential assessment, which really means that you take a walk around the open workspace – we usually work in such a workspace – and check where people are the most dissatisfied with the things they do at work. The assessment of potential takes different forms.
In the classic model, during the assessment some consultants walk around the office and within, say, up to a dozen or so weeks they have a map of all processes in the business. As you can see, in such a model this wouldn’t take 3,5 weeks. The advantage of this approach is that at the end we have a map of processes. However, unfortunately, the disadvantage is that we get this map very late. And sometimes at the end of the analysis those first processes we checked don’t necessarily look the way they did at the beginning.
Our approach is a bit different. We check one, two, three areas where this potential is visible at least at first glance. We don’t ponder whether this is the best area or the most optimal or if it brings the most benefits. The essential measure is whether this area is really receptive to automation and whether the people there do tasks which they should have stopped doing a long time ago because the tasks don’t bring them any satisfaction or inspiration, but rather they are some kind of unpleasant duty. After a process is selected, the assessment of potential ends within the first week. In the case of hapiloans it took 2 days, which we spent here with people who handle the first process.
After that we had the first good candidate for robotization. Then some boring things happened, which I won’t mention, until we started creating robots. The people responsible for creating robots in our company are called Robo Shepherds. I’ll say more about them in a minute or two because it’s something – the idea and the people – we are probably the most proud of.
At the end of this process there is a pilot launch. After the robots pass their tests and they are ready to enter the environment where employees of hapiloans work, we have to make the final check if the robots do what they’re supposed to. We don’t immediately let the robots do the full scope of their duties because at the beginning perhaps we don’t necessarily trust them all, but primarily we want our clients to develop that trust. So at first the robot gets a small section of the whole task of an employee, and in that time the employees sit in front of their screens, hands raised, and watch as their tasks are being done for them. After a few days, sometimes a few hours, when the trust has been built, the robot actually gets its complete set of duties and from this moment on it starts doing the job it’s supposed to.
Here are two subjects I won’t focus on too much, either. They are vital and extremely important, especially in a financial environment, in which hapiloans work. Data safety and change management. As for data safety, since we don’t have a lot of time, I’ll just say that data is absolutely safe and sound. Firstly, because we use the world leader in robotic platforms designed specifically to work in very regulated and very strict environments. Secondly, we use a cloud delivered by one of the world market leaders, Microsoft. We’re fully compliant with GDPR. We have full permissions separation. We watch over everything by strictly seeing to duties in the process of making robots.
Change management is important because it’s not like the robots, once you make them, go to work every day without any need for training. We know that environments change. We encourage our clients to grow their business without considering us in their growth. We don’t want to be a burden, we want to help. Every time something changes in the process, it is the Robo Shepherds’ duty to retrain such a robot and bring it back to work.
Robo Shepherds are a new profession that we came up with and which we like to promote. It is a metaphor because we could have called them robot engineers or robot developers. But we called them Robo Shepherds because they are responsible not only for creating the robot, but, most of all, for making sure that it comes to work every day.
I won’t be getting into technicalities too much. We’ll very quickly pass on to what robots actually brought to hapiloans. We’ll start with what probably is the first key aspect in every organisation – the employees. How did they react?
Przemysław Kamiński: After we said that we would be implementing robots at the call centre, there was some uncertainty. Because what does it really mean? What does it mean for an employee who works with a headset, who assists customers by phone? Will I become redundant? But soon it turned out that both employees and robots were doing great. Because we implemented automation, robots at the call centre, employees can focus on what is the most important: conversation with the customer. Customer service.
All those boring, let’s be honest, repetitive, simple tasks which so far were done by people are assigned now to machines. So now the application is ready to be processed by a consultant, they have all the necessary information, the application is complete and around 20% of those simplest, the least demanding, but very time-consuming tasks were delegated to robots. The employees are very happy with this change.
Aleksandra Rzeszutko: I’ll just add that we managed to engage employees to such an extent that right now they, too, search for places where the next robots can be introduced to so that they can dedicate all their time to our customer.
Mariusz Pultyn: It isn’t surprising, then, that employees welcomed their digital colleagues to their teams. However, above all, we make all those changes for the customers of hapiloans.
Przemysław Kamiński: I don’t think the customers of hapiloans know that robots support processes, but from their point of view the advantage is that they receive their money faster. The indicator calculated from the moment when the application is completely approved to the moment when the customer receives the funds improved by around 15%.
Mariusz Pultyn: So, finally, the standard question. Was it worth it? Did it pay off?
Aleksandra Rzeszutko: Of course. As Przemek said, on the one hand, we find the model of paying for robots cost-effective. On the other hand, which is actually more important, so first of all, our employees are satisfied. We know that a happy employee means an effective employee. Secondly, we don’t see such a high number of errors because robots do repetitive tasks, they’re trained for it and they’re not humans. Which is why they don’t make mistakes. As a result, we observe increased effectiveness both of robots and our employees.
Przemysław Kamiński: I’d like to add that robots also work at night. When people are sleeping, machines prepare applications received during the night so that when at 8 a.m. our employees come to work, the documents are waiting for them and employees can contact customers. We observe fewer mistakes. Errors were basically eliminated where machines replaced people.
Customers benefit because they receive the funds faster. Employees are happy because we have lifted off them the burden of simple, boring, repetitive tasks. From the business point of view it is also cost-effective for us because a robot costs around half the salary of a human. As a company we are satisfied with the realisation of this project and we are looking for further areas which we could automate in cooperation with our partners and hire more robots.
Aleksandra Rzeszutko: Perhaps I’ll just add that, as Przemek mentioned, we operate in different countries, which means we also work in different time zones. So this is where robots really come in handy as they work when here it is night, but in Mexico or Australia it is day.
Mariusz Pultyn: Of course, that really is a nice addition to what robots do. That is all we prepared in terms of the informative part of the webinar. Thank you for your participation and I invite you to join us for our future webinars.
Aleksandra Rzeszutko: Best regards.
Przemysław Kamiński: Thank you.
Mariusz Pultyn: Thank you.