Klaudia Bednarczyk
Communications Expert, Organizer of Hack'a Token - Raiffeisen Tech Poland
What happens when 75 colleagues from Poland, Romania and Kosovo get 1.5 days to experiment, collaborate and turn real business challenges into AI-powered solutions? Hack’a Token, the 2nd International GenAI Raiffeisen Tech Hackathon, brought together 19 teams to find out.
On 16–17 November, teams worked from Warsaw, Bucharest and Pristina, as well as remotely. All 19 joined the Development Stream to build working applications using Agentic Engineering, agentic workflows and AI-assisted development.
This year’s edition went one step further. Alongside building working applications, participants explored Agentic Engineering, modern AI-driven development practices and RBI PDLC, while paying particular attention to responsible AI and token consumption.
The result was 1.5 intense days of coding, testing, learning and exchanging ideas across locations.
The teams — 10 from Poland, 5 from Kosovo and 4 from Romania — each chose one of seven business challenges and had 1.5 days to build a working solution.
Hack’a Token was not only about what teams could build. It was also about how they built it.
Participants were encouraged to explore Agentic Engineering and AI-assisted development while keeping a close eye on responsible AI and token consumption. The hackathon also provided an opportunity to experiment with RBI PDLC, the new AI development framework, and explore how it can support modern software development practices. AI coaches and technical experts and organisers helped with technical issues, token limits and the available tools.
Each team had five minutes to demo its solution to an international jury of AI and technical experts, followed by a two-minute Q&A. The jury assessed innovation, business value, technical implementation, responsible AI use, presentation quality and overall impact.
After 1.5 days of intensive development and a full round of team demos, the jury selected the projects that stood out most strongly in terms of innovation, business value, technical quality, responsible AI usage and overall execution.
How do you turn an AI-generated interface into an experience that works for its users? For the five designers behind Design is my passion, the My HR challenge brought AI-assisted development together with UX expertise and user testing. They built a working application for employees, managers and HR, covering leave requests, personal data changes, payslips, contracts and training, with requests moving between the relevant roles for approval. Alongside implementing an AI assistant, they refined the AI-generated interface through UX expertise and user testing.
At the heart of LGTM’s prototype was RTECH: an AI banking assistant that can answer questions using the customer’s banking context — and refuse requests for investment advice within its defined guardrails. Around it, the team built a new digital banking application covering accounts, transactions, transfers, notifications and customer profiles. Automated testing, security scanning and CI processes supported the prototype, bringing the story beyond the interface to the engineering behind it.
Everyday banking was only part of the picture for AI Need Coffee. Their new e-banking experience paired a responsive prototype covering accounts, transactions, payments, cards, statements and other everyday banking services with an AI Financial Coach. The coach analyses a user’s financial situation to provide a financial health score, insights, risk alerts and recommendations. An additional AI assistant supports banking activities, bringing day-to-day tasks and financial insights into the same experience.
An answer is more useful when you can check where it came from. UnauthorizedAccess put that principle into practice in the Raiffeisen Knowledge Base Centre challenge, building an AI assistant that searches connected company knowledge sources and provides answers backed by citations. Users can follow those citations to original sources such as Jira or GitHub, ask follow-up questions and return to previous conversations. Access has a clear boundary, too: the solution respects existing user permissions, preventing users from seeing information they are not authorised to access.
MADS & Machine set its sights on Sprint Guard, designed to help Product Owners and Scrum Masters keep sprints on track. Their solution checks whether stories are ready for development, identifying issues such as missing estimates, insufficient descriptions or acceptance criteria, and missing Zephyr tests. AI can suggest improvements, while Scope Guard monitors changes introduced during an active sprint.
How do you turn an AI-generated interface into an experience that works for its users? For the five designers behind Design is my passion, the My HR challenge brought AI-assisted development together with UX expertise and user testing. They built a working application for employees, managers and HR, covering leave requests, personal data changes, payslips, contracts and training, with requests moving between the relevant roles for approval. Alongside implementing an AI assistant, they refined the AI-generated interface through UX expertise and user testing.
The next opportunity to learn will be an AI-focused event, conference or workshop in 2027. For selected projects, the journey may also continue through further presentations to business and technology stakeholders to explore development opportunities.
Hackathons at Raiffeisen Tech are not simply about prizes or rankings. They create a chance to experiment with technologies that are rapidly changing software development, test ideas in practice and learn from colleagues working across different locations and areas of expertise.
Beyond the competition, Hack’a Token gave colleagues across three countries a chance to test new approaches to AI-assisted development and learn from one another — experience they can take back to their everyday work.
Thank you to all participants, AI coaches, jury members, experts and organisers who made the second edition of our International GenAI Raiffeisen Tech Hackathon possible.
Want to get a feel for Hack’a Token? Hit play and meet the people who brought the ideas, teamwork and energy to life.
Communications Expert, Organizer of Hack'a Token - Raiffeisen Tech Poland
People, Culture & Communication Lead, Organizer of Hack'a Token - Raiffeisen Tech Romania