life at innovatricsback to all articles
From Fingers to Faces. From Technology to a Whole Product.
How has Innovatrics evolved from providing clients an engine to equipping the entire “car” – including a gearbox, training on how to drive the system, as well as subsequent servicing? MATÚŠ KAPUSTA, Innovatrics Head of Government Solutions, describes the development of the company, which initially supplied only the technology.
Being a biometrics company, you started with fingerprints, and now, you’ve expanded to faces. Tell us more about that.
“Our first product was the Fingera attendance system. At the beginning, we only had the technology capable of comparing fingerprints. However, it was not a self-sufficient solution. It was like making a car engine, but without the chassis. Not even a gearbox. Because we wanted to move forward, we needed to package our technology into something. We needed a whole product.”
That seems straightforward. You’ve managed to develop a complete product. How did that come about?
“It was quite daunting. We wanted to sell Fingera in the Czech Republic since I knew the market and had already worked in Prague before. After about a year and a half, when it was clear that we were only incurring losses in the Czech Republic and that only our Slovak operation was profitable, we had to strategize about our future. That is when we decided that we had to make not only the technology, but the whole product. At the time, we only offered the attendance system. Our other technologies required packaging too. Simply put, we developed software that clients could use immediately. We started to hire new people to program the applications.”
Is that what drove you to move forward?
“Thanks to that, we were able to make whole products and present them at exhibitions, where we we looked for potential partners. However, we needed a demo to showcase the technology. That is how we succeeded in winning a big project for the elections in Burkina Faso in 2011. Compared to our other projects, we delivered a complete solution there. Not only was the technology integrated into the client’s solution, but our complete application as well. It was the first time we sold a product that came with services – training the staff and operating the system. Gradually, we transitioned from a technological supplier to a complete solution supplier, and the market followed suit. We hired new people and the company started to grow organically.”
Was it a universal solution or tailor-made for the client?
“It was a solution that could be used anywhere, but customised for the elections in Burkina Faso. Through such a reference, we have won other projects as a solution supplier. In some cases, we even supplied more technology and less products. Sometimes more products and less technology depending on the nature of the project. However, all those projects were still based on fingerprints.”
Today, your business is largely based on facial biometrics.
“With the development of the neuron networks sector, more pressure has been placed on facial recognition. Sometime around 2014, we also started with facial recognition. Our first success was the evaluation through standardisation of whether a face has the correct parameters. This means, for example, that a person should not have a shadow across half of their face and is looking directly into the camera, and so on. We sold the solution to a client, but we wanted to go even further – developing technology to compare people based on facial photos. We wanted to replicate our success in fingerprints. For a long time, however, we were not been able to find skilled people who could handle the development of this technology at a high level. In Slovakia, we did not find anybody at all.”
What steps were taken after?
“We found two people in Brno, where we subsequently opened another branch. Interestingly, one of them, who has been managing the business in Brno, is Slovak. Presently, they also deal with iris recognition and other projects such as computer vision (image recognition). We have expanded in the field of technology as well as products. By constantly improving our technology, our products have stayed in demand. Ultimately, how well we sell depends on the product. It can ruin the technology when it is subpar.”
Has something like this already happened?
“We have a demanding client in Saudi Arabia – it took us a year from the first delivery to final acceptance until this client signed a contract with us.”
How did you overcome that situation?
“They wanted a system that would be better than the then-existing one. We were better at some things, but not at everything. We worked hard on it and adjusted everything according to their requirements. Thanks to that, our cooperation has been very smooth. In Saudi Arabia, we have moved even further than before. The client ordered a complete solution from us together with follow-up support. We are available 24/7 for the client, on call within 20 minutes. This means that we need to have our person there all the time as the system is offline for security reasons. We again had to hire more people rotating in about two-week cycles although our colleague has already been there for three months now due to the pandemic. This contract represents our continued progress.”
Have you encountered a challenging request from a client?
“Half a year ago, we dealt with the elections in Guinea. We worked on this project completely by ourselves. Prior to that, we had started every project with a partner where we were mostly just a subcontractor. However, based on our good results in Burkina Faso, we were given the opportunity to supply the complete system – software for 4,000 electoral registration stations. We did the data collection, back-up, biometrics. Finally, we printed the electoral rolls. It was an enormous challenge especially since it was our first. Owing to our strong technological background, we were able to make an algorithm at the client’s request, which identified up to 20,000 voters as minors. In Guinea, there was electoral fraud by registering kids with false documents. The elections took place there in March just before the pandemic.”
During this Coronavirus pandemic, have you identified any market gaps?
“Due to the pandemic, people are suddenly not willing to use fingerprints. Before it, our DOT (Digital Onboard Toolkit) product was mainly ordered by technology companies or major banks. Now, everyone wants it. It is a product that can remotely verify that it is really you – by scanning an identity card and your face. For example, large banks use it to remotely open an account via internet banking. Currently, we have been figuring out how to get this product into a “box” to make it more available even for less lucrative segments. This is the trend today because everyone wants to minimize physical visits. The pandemic has incredibly accelerated its development. Fortunately, we have already diversified our portfolio.”
Does getting the DOT product in a “box” mean that you do not have to provide support?
“Basically, yes. Large companies and banks do not care. They can afford to pay for training. But when a company that registers accommodation providers has to pay for it, it cannot afford to pay for 20 days of training, nor does it have an army of developers. In each case, our new market is precisely where the identity of an individual – whether a customer or a service provider – needs to be verified. This is where we need to offer a comprehensive solution as the market is incredibly competitive in this segment.”
Speaking of strong competition, how do companies stand out and how do clients choose their suppliers?
“There is strong competition among companies that deal with facial biometrics. In the field of fingerprints, it is stable as there are six companies worldwide at the top level. Being one of them, we have also succeeded with faces. Last year, we were in first place in the world for a while thanks to our algorithm, which was the most accurate among more than a hundred others. A company that wants to be one of the leaders must provide bug-free solutions. Identity verification must be above 99 percent. One hundred percent is, as of yet, an unachievable goal. One other very important thing is vitality verification. That is, whether a live person was photographed or someone just stole your photo from the Internet and wants to register under your identity.”
This should work in the case of smart quarantine, as well.
“Yes, it is part of our solutions.”
How are you doing competitively?
“Today, we are about 15th place with our six-month-old algorithm. But to stay at the forefront – among the top 200 companies – we have to improve it approximately once every six months. If we did nothing, we would move down to 180th place and probably our clientèle would also decrease. We have to show that we are able to keep up with the times. Thanks to the US Government, which tests the algorithms, small companies can easily compare themselves to large ones. Also, the our company profile has been raised through those tests.”
Is it possible to go even further in biometrics?
“Certainly yes, we can have better, faster, more accurate technology. Our goal is to have the highest possible accuracy regardless if it’s videos or photos. By achieving that, we are able to provide quality solutions in searching for people even with a low-quality photo. This is what is measured in benchmarks. The best ones have a search accuracy of 99.7 percent; the worse ones 99.3 percent; and it has continued to improve. The future is heading in that direction.”
See all vacancies