SAS is the founder and future of analytics and a world champion for Responsible AI and Data Ethics. In the 16 years of its presence in Serbia, SAS enables data-driven digital transformation for public agencies and business organisations, while actively promoting digital literacy as an integral part of human-centered innovation
Here SAS Adriatic General Manager Mrs Milatović Skorić explains for CorD Magazine how to scale human observation and decision- making, and how to leverage automation to enable the kind of real-time decision making that’s essential in the digitalised modern world.
In which industries and areas of life do you have the strongest presence? Where is it easiest to see your contribution?
– Since it was founded in 1976, the mission of SAS has been to help governments and businesses make timely and informed data-based decisions. The need for the optimisation of decision-making processes is recognised in many areas of the Serbian Economy – in financial services, telecommunications, retail & supply chain management and energy, but also in public agencies.
Promoting intelligent decision-making is more important today than ever before, as we are now creating data that is outpacing human capacities. This is where Artificial Intelligence becomes an integral part of the future of our businesses and our communities, for data preparation, automating modelling processes and routine tasks, for models to cope with the increased volume of small-scale decisions that are necessary due to customers’ expectations for real-time transactions and personalised offerings.
Research shows that only 35 per cent of companies apply analytical models fully in their operations. Why is this percentage so small?
– If we measure AI’s current effectiveness in terms of accurate and scaled decision-making, then I’m afraid the research you cite is correct: more than half of the models developed never make it to production; and 90% of those that do take more than three months to deploy, while 40% take more than seven months.
The cloud is the missing link in the AI value chain for organisations, for scaling decision-making and human observation.
Enabling wider and better applications of language technologies for the Serbian language is vital for the country to fully leverage modern AI capabilities
However, in order for the cloud to fulfil its potential, organisations need to think outside the box of the “cost reduction” imperative that often accompanies decisions regarding cloud-based operations and begin considering ways that cloud-based analytics can provide them with a competitive advantage, in terms of agility and resilience.
At the same time, it is important for organisations to adopt technological solutions that support innovation ecosystems across cloud providers, data formats, and even full open-source integration.
How important is it to know how to collect, interpret, understand and utilise data?
– Data literacy is a fundamental concept for companies and communities to pursue human- centred innovation through AI. In order for people to trust AI, they must understand AI and data. In this notion, we actively promote data literacy among Serbian people through strategic partnerships, such as the one we’ve forged with the University of Belgrade. At the same time, for countries like Serbia, it is important to create the conditions that would permit our nation to fully leverage AI capabilities. For example, large language models, such as OpenAI GPT, have great capabilities for natural language processing.
However, they are not well adapted to the Serbian language, and even less to the specific uses of our language (like those present in law, medicine, finance and media). That’s why we are proud to support the NLP project that was initiated by Slobodan Marković from UNDP, Vuk Batanović from ICEF and Dr Tanja Samardžić and that will be implemented by the Innovation Centre of the School of Electrical Engineering, University of Belgrade, and ReLDI (Regional Linguistic Data Initiative). This project aims to create quality data sets and AI models adapted to that data, which would be publicly available to everyone under permissive licenses, thus creating space for wider and better applications of language technologies for the Serbian language.