#5 – How we created Facial Recognition Payments at OP

How did our team of experts create Facial Recognition Payments at OP?

In this 5th episode of OP Tech Podcast our host Kristian Luoma interviews Technology Expert Thomas O’Rourke. Together with his team, Thomas has developed a whole new transaction method at OP: facial recognition payments.

In this episode, Kristian and Thomas talk about what are the facial recognition payments, how the actual transaction process was developed, and what are the advantages of using Elixir. Elixir is a programming language for developing a backend server for facial recognition payments.

Remember to check what our last episode (in Finnish) was about >>

What are facial recognition payments, and how were they developed at OP?

“As its core, facial recognition payments mean you show up, show your face and buy something,” Thomas explains and continues:

“It’s about the ease of transaction: When you’re buying a coffee, you walk up to the cash register – and in the good old days, the cashier knows who you are. In this case, we’re using a bit of sophistication on the backend to know who you are and then charge you for it. It’s easy, fast, and pretty cool.”

The facial recognition payments development process started with the ending line and setting expectations: what the team wanted the transaction to feel like for an end-user.

After a few prototypes, Thomas and his colleagues started with developing the backend server part in Elixir, putting streams of data to the test right away. The agileness of the team indeed came out when the group compared different recognition engines and understood advantages and disadvantages. At the same time, they did a lot of user testing to figure out what kind of process would work the best. 

“We had much confusion on our first try with a user first showing a QR code from a phone. It was not intuitive at all. We turned our architecture around in a few days, and the second factor was a lot better. When we integrated with push messages, we were capable of doing the whole payment process within just 3-4 seconds,” Thomas says.

Elixir inspires Thomas

One of Thomas’ favorite technologies is Elixir; a language used when developing a backend server for facial recognition payments. Compared to most languages, Elixir has its runtime and is very sophisticated.

“It’s one of the only systems that can give you soft real-time performance. It allows you to create processes and do things in parallel without having to consider artificial matters. I can have ten thousand processes, and it’s still very light,” Thomas says. 

“And it has never crashed,” he continues.

Elixir works great when managing multiple background processes. Surely sometimes things can blow up if, for example, one sends terrible data to the server. But still, the situation is easy to fix.

“Something can crash inside the server, but then the appropriate routines pick it up and restore the state, so with Elixir, you don’t ever have to restart the whole system.”

Listen to the whole interview on Spotify >>

Episode content:

  • What does facial recognition payments mean?
  • How did Thomas and his colleagues create OP Facial Recognition Payments?
  • What’s the value for customers?
  • The advantages of using Elixir
  • Why Thomas enjoy working at OP?

Episode time: 19:34 min

Let’s continue the discussion online! Connect with Kristian and Thomas on