DOI: 10.33808/clinexphealthsci.1990829 ISSN: 2459-1459

Development and Evaluation of EczaciGPT: An AI-Powered Chatbot for Medication Information Retrieval in Pharmacy Practice

Nazlı Tokatlı, Muhammed Yunus Bektay, Fikret Vehbi İzzettin
Show PDF Cite
Objectives: EczaciGPT is an innovative AI-powered chatbot designed to assist pharmacists in accessing accurate, timely, and comprehensive medication-related information. Leveraging advanced large language models (LLMs), it provides detailed responses to queries regarding drug interactions, side effects, dosages, contraindications, and more through a user-friendly web interface.Methods: The system’s architecture is composed of three primary layers: a user interface for seamless interaction, a robust server for efficient query processing, and a data layer containing an extensive repository of drug-related information. EczaciGPT also integrates with external resources and databases to ensure up-to-date and reliable outputs. Built on the Dante AI platform with GPT-4o Mini (OpenAI) as the underlying large language model, it supports pharmacy workflows by offering dependable insights from medication-related documents, guidelines, and research literature, ultimately streamlining information retrieval and enabling pharmacists to address patient concerns more effectively. A custom online questionnaire covering four key domains (user experience, usage patterns, relevance and accuracy of the information provided, and technical performance) was completed by 21 pharmacy students who used EczaciGPT during their practical work.Results: All 21 participants (100%) reported ease of communication with the tool. All rated the information as “very understandable” (15; 71.4%) or “understandable” (6; 28.6%), and as “definitely correct” (10; 47.6%) or “correct” (11; 52.4%). Twenty participants (95.2%) agreed that the tool reduced the time needed to understand and apply information.Conclusion: These results suggest that pharmacy students perceived EczaciGPT as easy to use, clear and accurate. Objective validation of response accuracy and evaluation among practising pharmacists are needed before conclusions can be drawn about its impact on pharmacy workflow and patient care.

More from our Archive