Bridging the gap between developers and algorithms

Croatian startup airt is pioneering AI technologies and moving fast, proposing urgent incentives for investors and startups to unleash competitiveness in global markets.

airt is a Croatian AI startup with a mission to use artificial intelligence on selected data ‘in the simplest, easiest and quickest way’. It was co-founded in 2019 by Davor Runje and Hajdi Ćenan. 

In less than three years, they introduced fintech solutions for small and medium banks. They applied for the global patent of their deep learning algorithm while being prominent members and ambassadors of the Association of Croatian Independent Software Exporters CISEX. 

A few days ago, at the Infobip Shift 2021 conference, airt introduced a private beta of their platform. 

Hajdi and Davor are a well-coordinated duo who started their collaboration in Drap, the first Croatian digital agency. As technology is their field of research and play, they joined forces. 

Hajdi brought her knowledge of media and marketing to their collaboration. Davor has a programmer and scientist background and is one of those serial entrepreneurs always on the lookout for the next big idea. 

airt is one of the pioneers of the AI industry in Croatia, intending to become globally recognised. If we have learned anything from their past ventures, we expect nothing less this time either.

We asked them a bunch of questions and Hajdi was the one helping with our curiosity. 

Why did you decide to focus on artificial intelligence and deep learning? 

Davor is always on the lookout for new technologies; what’s the next thing that will mark our lives in the next period? 

About five years ago, there was a lot of buzz around several things simultaneously: AI, blockchain, augmented reality (AR) and virtual reality (VR), and the internet of things (IoT). After some thought, AI seemed the right direction for us. 

There was also a desire to change direction, as digital marketing has become a commodity and did not present a challenge and opportunity for personal growth. 

Artificial intelligence proved to be a technology that will dominate the next few decades. It is being equated to the Internet, electricity even, as it will transform our personal and professional lives. We’re already using it daily, even though many people are still not aware of that. 

But every time you scroll through your social media timeline, shop on Amazon or search Netflix, there’s an algorithm working in the background to personalise or recommend what’s best for you. 

These examples are still mainly in the domain of the most prominent players, but very soon small businesses will have to start championing AI, and we know this as we’re working on making it happen.

Why airt choose the banking sector first? 

Our work kicked off as a consulting project in one bank and grew into building an AI platform once we realised the full depth and width of the problem and the potential of its solution. 

Many companies, particularly those in the financial and communication industries sit on large volumes of data but use only a tiny fraction of it when building predictive models. 

No wonder, since that is not their primary business, they also don’t have enough time or resources to develop that process in-house. 

How did the first banking project influence your next steps? 

We have spent almost three years deep-diving into event-based customer behaviour. That enabled us to innovate, so we submitted a patent application for our deep learning techniques for that type of data to extract the most value from it. 

Any business with event-based customer data can use our AI platform, even if they don’t have data scientists. 

What is the application of AI across industries? 

Companies are very aware that they need to implement AI to compete and sustain their business, but implementation is still a huge problem. 

Almost half of AI projects never go from the experimentation phase into production. Another third fails due to a lack of production-ready data, integrated development environments, and necessary expertise. 

Data scientists are already extremely scarce, especially those with domain knowledge. Even large companies cannot find them, so what chance do smaller ones have? 

Developers are sought after, but there are many more of them; however, they are not necessarily familiar with new technologies such as AI.

Have you found a solution to that problem? 

While working with a large SaaS vendor, we realised we’re ideally positioned to build an AI bridge for developers. 

We already spent so much time building an AI platform and techniques that can show results in a matter of days, so why not simplify the implementation of our know-how? 

Our platform is a low-code tool that can be implemented and used with just a few lines of code and some help from APIs. 

Developers don’t have to learn new languages and frameworks; they can let the platform do its magic in the background and provide the prediction results to act proactively, not reactively, towards their customers. 

Conveniently, last week at Infobip Shift, we announced that our platform is now available as a service and tailored to in the private beta at first. 

What inspired the event-driven route? 

We already live in a complex and fast-changing world. Following that environment, customers’ expectations are advancing too, which means businesses need AI, and it’s no longer just something nice to have. 

For example, if a business wants to know which customers might churn, they can use our platform to predict which ones are most likely to do so, react to this before it happens with a personalised note or offer, and keep them. 

Maybe they want to figure out which customers are most likely to buy a particular product; by offering it to select customers only, they will save a lot of time on contacting only those interested and maximise sales conversion. 

Simply by knowing beforehand who should be contacted and offered that product. 

Naturally, every business has different data, customers, and needs, but our platform is flexible enough to predict any combination of events or non-events. 

So, we might’ve started with banking but have expanded to any industry that has event-based data and wants to predict the behaviour of their customers. And now we want to make it accessible to as many businesses as possible, no matter how big or small.

Writing a patent application is a long game of patience and a tiresome and expensive process that took us almost six months to complete. The review of the application by the European Patent Office (EPO) is another lengthy process that can take up to 12 months before we even know whether it will be approved or we’ll have to make any modifications or additions. However, it’s a game worth playing. 

How does it match your goals? 

It’s part of our business strategy to create an Intellectual Property portfolio and to license our technology. 

The IP strategy also adds to our startup’s valuation. It paints our current technological and business model and its future and sends out a message that our startup does have a new and innovative approach and concept that has potential and gives us a competitive edge. 

That competitive advantage is another reason why we went down this road. 

We’ve invested a lot of time and money in building the innovative solution that we can now offer to the market. 

What’s next? 

That solution is already better than the competition – IBM’s TabFormer. Our approach also means that we will continue to innovate and build the IP portfolio further. 

We want to be strategic about using our IP, from licensing our innovative solution to licensing parts of it to other tech providers. 

We, therefore, already have additionally connected innovations in our pipeline. This, on the one hand, requires relatively high expertise of our team and high financial investments, but on the other hand, does not limit our growth potential. 

In other words, our market is the whole world. 

How do challenges with a lack of governance framework, locally and globally impact your everyday work? 

Entrepreneurship in Croatia is not an easy task, predominantly due to the amount of red tape. Creating a startup with global ambitions is even more challenging. 

Namely, startups can be differentiated from ‘regular’ businesses by building a (mostly) tech-based solution but, more importantly, can be quickly scaled globally. 

Additionally, startups are businesses that will usually not generate revenue from day one; on the contrary, they will have to invest in research and development before creating a solution that will see the light of day. 

This means that startups need various financing opportunities from day one, maybe even sooner.

When it comes to governance, which areas would you say are of highest priority? 

The investment scene in Croatia is still young and does not provide many options. Unfortunately, the situation is not much better throughout the rest of the EU, which is why most startups end up relocating to the UK or the US. 

If we want a thriving and fostering startup ecosystem, this is one issue that needs to be solved.

The good news is, there’s no need to reinvent the wheel, as we have an excellent example. Namely, the UK has the highest number of unicorns in Europe, startups with a valuation of USD 1 billion. 

The UK has investment schemes that offer significant tax incentives to those that invest in startups, which significantly mitigate the associated risk, increasing the number of investments and, consequently, the number of startups created. 

The larger the pool, the higher the number of startups that might become new unicorns. This works well with angel investors, particularly through Seed Enterprise Investment Scheme (SIES) and immensely helps startups in their initial, most vulnerable stages of development.

Furthermore, these angel investors are private individuals who get an opportunity to diversify their investments and use their assets to help build the economy. 

Instead of that money ‘sitting’ in bank accounts or being used to buy real estate, it’s pumped back into the economy by predominantly being spent on (high) salaries. In short, such an approach helps foster innovation and creates new jobs, not to mention it raises the development and maturity of the economy through enabling projects that ensure and introduce the implementation of new technologies following global trends.