Institutional Repository of Vilnius University
Dirbtinio intelekto taikymas UX analizės automatizavimui, siekiant gerinti vartotojo sąsajos veiksmingumą /
Abstract
dc:descriptionThe aim of this bachelor's thesis is to analyze and implement an artificial intelligence-based solution for automating UX analysis to improve user interface effectiveness. The thesis addresses the problem of how to apply artificial intelligence technologies to automate user interface effectiveness evaluation while reducing the need for expert knowledge and time. To achieve this aim, seven objectives were set: to analyze scientific literature and summarize UX evaluation method principles; to conduct a comparative analysis of the most popular UX evaluation tools; to conduct qualitative semi-structured interviews with specialists; to develop the solution specification; to design the system architecture using UML notations; to implement a functional software prototype; and to test the accuracy of functional and non-functional requirements. The following research methods were applied: scientific literature analysis, comparative analysis of market tools, qualitative semi-structured interviews, and system design using UML notation. The literature analysis covered J. Nielsen's usability heuristics, R. Cialdini's persuasion principles, and WCAG 2.1 accessibility guidelines. Comparative analysis was conducted for six market tools (Crazy Egg, FullStory, Attention Insight, UXCam, Maze, Contentsquare) across 12 criteria. Five representatives from different professional groups participated in the qualitative research. During the design phase these diagrams were created: use case, activity, sequence, entity-relationship, component, deployment, and state diagrams. A three-tier architecture was designed with a separate browser service and a three-stage AI analysis pipeline integrating Claude Haiku, GPT-5.2, and Claude Opus models. A system of 29 heuristics was developed, organized into five impact areas: conversion, trust, engagement, usability, and accessibility. Testing results confirmed that 11/12 functional requirements were successfully implemented, average analysis duration is 1 min. 8 sec. (specified limit <5 min.). The results of the thesis confirm that the problem was successfully solved – the developed system enables automation of the UX analysis process, reducing the need for expert knowledge and time from several hours to almost one minute.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Andriuškevičius, Jonas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD277691972&prefLang=en_US