Table of Content
INTERVIEWS
How We Created the Atfinity Platform: Interview with Thorben Croisé
Ever since Atfinity’s earliest beginnings, Thorben Croisé was there to set the course and help establish the company. He is co-founder, CTO and a constant learner. Because in his view, one needs to learn more in order to do more. Thorben talked in-depth about Atfinity, from the time the company was just an idea, right up to today and the big projects we are doing for our clients.
How did the Atfinity story start?
Thorben: Actually, in the beginning, Atfinity did something completely different! We wanted to offer something for wealth managers and banks to manage a conversation between a client and the bank. A new, improved way for banks to chat with customers.
When we started offering this and having meetings with prospects, we noticed a pattern: they couldn’t use such software due to compliance reasons. This made us see this huge gap in the market; that, basically, every bank and financial institution was unhappy with how their compliance is handled. It was, and partly still is, a huge problem.
We talked to dozens of potential clients about what they did and what would help the most and iteratively got closer to something they really needed. We discovered that knowing what questions to ask a client (when onboarding, updating or reviewing) is complicated and it’s difficult to orchestrate all the existing tools and all the other providers you need or want.
So, if you need a signature, identification or a risk check, it’s complicated to orchestrate and figure out what questions to ask. What about a client’s Social Security number? When do we need that? Or if it’s a loan, what’s their income, where do they work? And if you need a more in-depth KYC, where do they get the money from?
So, we built a platform to combine Orchestration and Question Finding and have an intuitive user interface alongside.
What was the biggest challenge in creating an original solution?
Thorben: When we built our (AI-inspired) rule engine, in the beginning, we thought: okay, it’s a rule engine, that’s a problem that has been solved. Let’s just take a rule engine, ideally an open-source one, use it, and then we’re good. It turned out that we wanted to ask the right questions based on the rules. So, it’s a decision engine.
It’s a bit like having a conversation with a doctor. They ask you: What’s your problem, where does it hurt? And then you get closer and closer and find out what you have. And we did the same thing; asking question after question, to figure out all the information until finally you say: here’s what you need – your contract.
And it turns out that in the 70s and 80s that was a very active research area, and then it stopped. It’s like an earlier approach to doing AI, but then people have not researched that area too much. Because nobody found a killer use-case for it. There are applications but they tend to be special and very industrial. So, looking at the large open source community, this fell a little bit out of fashion.
In addition to that, even the general rule engine fell out of fashion. When I started studying (computer science), in the late 2000s, that was still something that people were using. And then ten years later, it was not. So both things, the decision engine (or expert system) that’s a very industrial solution and the open source rule engine were just not available. Which meant we had to start from scratch.
We had to do this tremendous task of building another rule engine, and not just any, but one that also calculates an optimal flow chart. Every time I talked to colleagues, they all thought it’s potentially a bit crazy that we’re building this.
Fortunately, at that time I actually didn’t talk with too many of my colleagues and I didn’t know it was crazy. So we just started working on it. And that’s sometimes not a bad idea, to just do something nobody told you was impossible. Because here, it turns out it was possible!
In a bit less than a year, we built such a rule engine. It was completely novel (never been built that way). I believe that nobody has ever done this, to create and combine a decision engine (expert system), a rule engine, AI and its own rule language around. And then to also have a user interface, that’s a complete novel solution out of four, maybe five different pieces. They have all existed, at least in theory. But in practice, put together, it was completely new. That made the technical aspect very challenging.
It sounds like it’s just another rule engine, but it’s not. It’s a completely different approach. It’s a rule engine married with AI, married with process orchestration, married with its own rule language. It’s huge. And every time I explain to people what we built there, they all think: Okay, that’s a lot, why did you do it? And my honest answer is that I didn’t know it was impossible. Nobody told us that it would be hard.
We worked at it for years and put all these parts together. And now we have this tremendously cool piece of software that combines all these things.
What makes Atfinity’s platform different from other solutions?
Thorben: What makes Atfinity unique and different is the combination of so many complex tools.
First, you will find rule engines that mostly work in the background. Large providers will give you a rule engine. However, these rule engines have no user interface (when they run). So when an event happens, and based on that you need to ask a question – if you have no user interface, how do you do that? That happens actually quite often in business processes. We offer a rule engine with a user interface to execute the rules. That’s already very unique.
However, that’s not all. There are a lot of these rule engines or automation tools, and they are meant for developers. So if you’re a large bank and you have 10,000 employees, it makes sense that you have a team of, let’s say, five developers. It helps you a lot to actually take a rule engine, make it ultra efficient and program it with code. Okay, fine, but what if you go no-code ?
We thought: We want this rule engine to be programmed by people that have no programming experience. That’s what Atfinity offers: You don’t need an engineer to create an efficient, sustainable software solution. This makes it even more different.
There are no-code solutions that could do parts of this. But a no-code solution that has an AI rule engine, a user interface, allows you to orchestrate your processes ( and make API calls in between) and connect to an identification engine and the signature system? The combination of all these things is completely unique.
This combination is what makes it so powerful. If you need a software development team to configure Atfinity, the flexibility would not be there. With us, you just need the domain expert, maybe a compliance officer or a legal expert, and they can configure Atfinity. That’s unique!
We give you this interface for your users (relationship managers, compliance offers, risk managers, board members, etc.) directly out of the box. And as a domain expert, you can tailor the user interface to make sense to your company, fintech, bank, or neobank. That’s also very powerful.
How exactly does Atfinity facilitate process orchestration, because it’s very complicated to have everything aligned?
Thorben: First, maybe let’s see what process orchestration is. If you have a process that just means 10 data points, with no rules and nothing else, then you would not really need to orchestrate. A lot of the need comes from the fact that, for any modern business process, there are many tools and stakeholders involved.
If you’re onboarding an employee, you need to put the employee into the HR system and into the payroll system. If you’re creating a bank account, you need to put client information into the core banking system, into the CRM system, and you could need a credit check and maybe also get a digital signature… So, it’s four or five, ten, or maybe more systems we need to call at the right time and get the information from the system.
Process orchestration really means having an orchestra (of tools) in front of you. You have all these different systems in front of you, and you’re a conductor telling them: do this, do that, give me this information, etc.
However, if it would be an actual orchestra, you would have a music piece do A, B, C, and it’s completely fine (even required) to never change the order. But it’s not really that way for a business process. For example, you ask the client what they want and only then you realise the client wants to have a credit. Now you need to call your risk system. Or maybe get an extra approval process from the management, since the client wants a large credit.
The point is: after you do A, you find out it’s not B you need to do. You need to do F instead. While you’re conducting your “orchestra”, based on what other people play, you play something else.
So, it’s a bit like jazz, right? You react to your audience, you react to the mood, to what’s currently happening. In this sense, it’s not conducting an orchestra, it’s like conducting a huge jazz band in a techno club where you want to have people dancing all the time. You react to everything that’s happening around you. And that’s why Atfinity gives you rules for this.
The client tells you their income is a hundred thousand euros a year and then we need to make an assessment based on that. Next week your client comes and and says: Sorry I was wrong about it – it’s not exactly a hundred thousand, it’s only 92,000 euros per year. Then, we need to redo the check. Because every time the data changes you need to check with all relevant systems. Atfinity will do that automatically based on the simple rule you provided: Do a risk assessment for the income of a client.
Orchestration with Atfinity in general means we have rules that do certain things, talk to or call different systems. It’s like you’re saying: The violin only plays after the trumpet has played a C minor. And then we basically do that, and every time the trumpet does this, we do that. We react to the data, from your income data and your first name to your nationality, domicile, the way you told us what your KYC situation is, etc.
What do businesses lose by not automating their processes?
Thorben: I mean, business has changed: The world around you got a lot more complicated. The regulations are a lot harsher and there’s a lot more of them. Before, maybe you needed to answer only ten questions for specific information. Now, it’s a hundred. So, your costs are higher, and not automating means that suddenly onboarding clients is potentially more expensive.
For example, let’s say before it made sense to give credits from maybe EUR 50.000 upwards. But now, just figuring out whether you can get a credit or not is so expensive that below 500.000 it’s not worth it. But there’s probably a huge market between 50,000 and 500,000. Just because it’s not cost-effective for you doesn’t mean the market went away.
So, if you automate you get your clients back. Or, the other way around: it gives you new opportunities to enter different parts of the market. We see a lot of clients that, once they started automating with Atfinity, suddenly got access to a huge, new part of the market. They could go to clients that had less money or wanted smaller credits. Or go to a retail market and open up to a broader audience.
Also, it’s good to mention here that if you don’t automate it will potentially be very annoying to the people that work in your company / bank. Before, having ten questions was fine, but now it’s a hundred questions and everybody in your bank is a little unhappy. Because if you’re a client advisor you probably like advising clients; you like your job. You like clients and you like giving them a good service. But now you’re frustrated that it takes weeks to get to an answer and onboard them.
That frustration also means that they’re not as happy with their job as before. And since they are not as happy, they probably provide a worse service and aren’t as nice to clients. So, the service quality is lowered. Your clients don’t like that it takes long, your employees don’t like it. They provide a worse service, your clients like it even less… It’s a vicious circle. And your quality declines. That means, sooner or later, you’re about to lose business.
How can people find out more about Atfinity’s platform?
Thorben: The best way is to book a demo, for sure. I always say that talking about the product is one thing, but actually seeing in practice what it can do is something else. In the demo session, you can learn in-depth about our platform, its many features, and the benefits of implementing it in your organisation. Also, you can ask as many questions as you need, to clarify any grey areas and get the full grasp of Atfinity’s effectiveness.
Book your demo today and see why leading financial institutions
worldwide trust Atfinity to drive their digital transformation.
Book your demo today and see why leading financial institutions worldwide trust Atfinity to drive their digital transformation.