Working with the optimization of complex business processes eventually led Flowit to machine learning. Juhan Pukk stresses that machine learning is not about teaching a computer to visually identify pictures of cats or something similarly trivial. “One of our greatest breakthroughs came when we worked with logistics company DSV, who possess an endless array of data all collected into a single spreadsheet. With this data we developed a solution to help them approach clients with pre-filled offers before the client has even asked for one. This is how we ventured to the world of predictive analytics,” recalls Juhan, adding that while a salesman’s hunch or gut-feeling is a good character trait, concrete data provides far more superior input for customer engagement, leaving clients baffled about the unprompted yet remarkably accurate quotes that they didn’t even realise they needed.

Many have tried to harness machine learning for the benefit of their processes, but only a few have succeeded. “It’s a very scientific field, but therein lies the problem – science is always aiming for maximum reliability of predictions and won’t settle for less than 100%, but companies can gain a competitive edge already with a 70% accuracy in predictive analytics,” remarks Juhan, adding that science seems to aim for whatever the goal is, but it’s not necessarily the company’s ROI. “We need to find the balance between these two worlds and understand that even if there’s a 20% chance of risk of failure, it shouldn’t be regarded as a failed attempt in business terms. We at Flowit work closely together with universities in an attempt to reconcile the scientific and commercial approach to predictive analytics. We validate the client’s goals and needs before taking on the project and know exactly which algorithm goes where, thus reducing the risks even further.” Juhan admits that there are still quite a lot of myths surrounding artificial intelligence and machine learning. “First, there’s the expectation that AI is something that can independently discover a problem and immediately offer a solution. This is not a realistic expectation and technologically we are lightyears away from this level. Another myth is that machine learning is a technology that takes considerable time to implement. Busting this misconception actually landed us our first major client in Germany,” Juhan says.

The crowning achievement of Flowit’s expansion to the German market was the landing of vehicle
manufacturer Daimler as a client. The German industrial giant was unhappy with most machine earning software projects taking 1-2 years to implement. Along came Flowit and managed to reduce the implementation cycle to mere 2 months, helping the carmaker maintain its innovative edge. Besides machine learning and predictive analytics, Flowit’s core competence lies in the development of self-service environments in addition to machine learning and predictive analytics. “Today we see a number of companies paying more attention to user interfaces and the overall user experience. This is why our trademark is our focus on usability. 15-20% of our software product developing time is spent on machine learning, but the remaining 80% is still all about simplifying processes and making the product usable for the customers and keeping it as simple as necessary,” explains Juhan.

Large corporations are used to operating within closed-box software solutions that leave little room for additional tweaking or development. Flowit strikes a difference here by basing their software products on open source platforms and libraries which allow a greater degree of flexibility to adjust the final product to match the customer’s specific needs or future upgrades. “But without making any compromises on security – our software’s security protocols are on the same level as those of major banks. Security may inhibit innovation, but there are always creative ways to work around such constraints,” believes Juhan. Another key advantage for Flowit is the transparency of project budgets. “There are no hidden costs, the client always gets the product for the price that was quoted and the projects are always delivered on time. This is the cornerstone of our work ethic and we do not make compromises on this,” stresses Juhan.

Companies across all sectors are now taking more time to analyse their activities before proceeding with software projects and Juhan applauds this. “There’s less of a “let’s get it done now” attitude, instead there’s a drive for efficiency. Clients should always keep their attention on their business objectives and their return on investment when commissioning new software.” The more business executives become aware of technological possibilities, the more likely they will reach their targets, such as increasing revenue or reducing drop-offs in e-commerce solutions. “However, IT or digitalization is only a part of the toolbox, you can’t overestimate the importance of marketing and branding and sales to drive your
business forward. Keeping the solutions simple can often be more complex and time consuming that creating a digital behemoth that requires months and months of training. Our systems are designed in a manner that only requires a 10-minute online tutorial to get the users acquainted with the software.”

Although success on export markets always requires a lot of effort and planning, Flowit has literally managed to stumble on some remarkable client wins. “We have taken the German market to our hearts and have worked closely with Enterprise Estonia to be present there. The German market is incredibly attractive – a tough market to enter, but impossible to leave, as we say. It’s important to be present, to interact and to gain trust through personal contacts.” One such stroke of luck hit Flowit at an AI seminar held at Hamburg’s ARIC club, where Juhan carries the distinguished title of ambassador. After his presentation in a panel discussion featuring Estonian government dignitaries and German entrepreneurs, Juhan accidentally slipped on the carpet and collided with a local businessman, who complimented Juhan on the honesty of his presentation. The businessman in question turned out to be a board member of the world’s largest microchip manufacturer Nexperia and Flowit has been helping them with software solutions ever since!

Other notable clients besides Daimler and Nexperia include Nasdaq, Telia and the chemical concern Linde Group. Juhan also serves as the president of the Estonian ICT Association and has a very good understanding of the strengths and unique skills of other Estonian IT companies. “I can only describe Estonian IT entrepreneurs and engineers as amazingly inventive and results-driven professionals, who are happy to think outside the box and work relentlessly to achieve the set goals.” Estonians have also been characterised as resilient and consistent business partners, who take great pride in brutally honest communication with customers. “With any given development project there are only three things that can go completely wrong – communication, communication and communication. It’s important to understand the client’s business and to think along on their level. Offering nothing else but just technical solutions is destined to fail, it’s crucial to see the big picture as well. If you want something done right and to keep it user-friendly – then welcome to Estonia!”