TETRA Designing user-friendly AI solutions

This project aims to accelerate AI adoption among end users through two complementary pillars. On the one hand, by investigating how AI interfaces should be designed to promote user trust and AI literacy among end users. On the other hand, by analysing how generative AI tools can be used responsibly in the design process itself. By combining scientific and practical insights with the requirements of the EU AI Act and privacy legislation, translating these into manageable guidelines and validating them through authentic practical cases at Flemish SMEs, an innovative approach is created that leads to increased conversion and a stronger market position for companies. This project delivers three concrete results that can be directly applied by Flemish SMEs: - a validated framework for user-friendly AI interfaces, consisting of practical design guidelines for three common types of AI interfaces o (1) conversational interfaces such as chatbots for customer support, o (2) data visualisations such as dashboards that make AI predictions transparent to end users o(3) adaptive and personalised interfaces that automatically adjust to the user's profile and skills, with a special focus on e-inclusion and privacy protection. Interactive demonstrators are being developed for each type. The framework also includes reusable UI components and interaction patterns (such as ways to provide feedback and display error messages), supplemented with evaluation criteria and test scenarios for user validation. The effectiveness of these guidelines is systematically tested through user research and practical implementations at companies in the learning network, enabling direct applicability. - a methodology for responsible AI integration in the design process. This consists of an evaluation framework for selecting suitable AI tools based on objective criteria (such as GDPR compliance and integration requirements), a step-by-step plan that guides companies from initial analysis to implementation, and practical checklists for managing privacy and security risks. - An accessible AI design guide for wide distribution.This guide combines the knowledge from both main pillars of the project: best practices for user-friendly AI interfaces and guidelines for responsible AI integration in the design process. The aim is to enable companies to apply general design principles and translate specific insights into their own context through documented and detailed use cases from Flemish SMEs, practical checklists and concrete examples. For each implementation, both the technical approach and the results achieved in terms of user acceptance and business impact are documented. - KPI 1: the number of unique companies and non-profit organisations identified during the project implementation that indicate they will apply the knowledge gained from the TETRA project: 30 and 75 after 1 and 2 years, respectively. - KPI 2: the number of documented validations that can be used to better disseminate the knowledge: 5 and 10 after 1 and 2 years, respectively. This includes detailed practical validations and workshop implementations. - KPI 3: the number of unique organisations consulting and/or downloading the online AI design guide: 50 and 250 after 1 and 2 years, respectively.- KPI 4: the number of students who will absorb knowledge and insights resulting from the project: 300 and 1,500 after 1 and 2 years, respectively. Application of the results The project results will be compiled in an interactive “AI design guide” that will be made available online. This guide combines practical design guidelines for user-friendly AI interfaces with concrete tools for the responsible use of AI tools in the design process. Validation through controlled testing, examples and practical cases guarantees broad and practical applicability.