Penerapan Metode Heuristic Evaluation dalam Analisis Perplexity

Perplexity AI merupakan chatbot berbasis Artificial Intelligence yang dirancang untuk membantu pengguna dalam pencarian informasi secara cepat dan akurat termasuk dalam konteks akademik. Meskipun penggunaannya di kalangan mahasiswa semakin meningkat, belum banyak dilakukan evaluasi terhadap kualitas kegunaan (usability) antarmuka sistem ini, penelitian ini menerapkan metode Heuristic Evaluation berdasarkan sepuluh prinsip Nielsen. Penelitian dilakukan secara kuantitatif dengan menyebarkan kuesioner online kepada 88 mahasiswa aktif Program Studi Sistem Informasi Universitas Malikussaleh yang telah menggunakan Perplexity AI. Data dianalisis melalui uji validitas, reliabilitas, regresi linier berganda, dan perhitungan severity rating. Seluruh item kuesioner dinyatakan valid dan reliabel (Cronbach’s Alpha > 0,6). Hasil regresi menunjukkan bahwa secara simultan semua aspek heuristic berpengaruh signifikan terhadap kepuasan pengguna (F = 88,89; signifikansinya 0,000; Adjusted R² = 90,9%). Namun, secara parsial hanya tiga aspek yang berpengaruh signifikan, yaitu Flexibility and Efficiency of Use, Help Users Recognize, Diagnose, and Recover from Errors, serta Help and Documentation. Hasil perhitungan severity rating menunjukkan sebagian besar masalah tergolong cosmetic problem, yaitu tidak wajib diperbaiki kecuali tersedia waktu tambahan. Rekomendasi perbaikan diberikan untuk peningkatan fitur fleksibilitas, bantuan sistem, dan dokumentasi agar Perplexity AI lebih adaptif terhadap kebutuhan pengguna. Hasil penelitian ini dapat menjadi rujukan dalam pengembangan sistem chatbot berbasis AI di bidang pendidikan, sekaligus memperkuat literatur akademik dalam kajian Interaksi Manusia dan Komputer (IMK), khususnya pada aspek evaluasi usability antarmuka digital.

Application of the Heuristic Evaluation Method in Perplexity Analysis

Abstract

Perplexity AI is an Artificial Intelligence–based chatbot designed to assist users in retrieving information quickly and accurately, including within academic contexts. Although its use among university students is increasing, limited evaluations have been conducted regarding the usability quality of its interface. This study applies the Heuristic Evaluation method based on Nielsen’s ten principles. The research employed a quantitative approach by distributing an online questionnaire to 88 active students of the Information Systems Program at Malikussaleh University who had experience using Perplexity AI. The data were analyzed through validity and reliability testing, multiple linear regression, and severity rating calculations. All questionnaire items were found to be valid and reliable (Cronbach’s Alpha > 0.6). Regression analysis indicated that, simultaneously, all heuristic aspects significantly influenced user satisfaction (F = 88.89; significance = 0.000; Adjusted R² = 90.9%). However, on a partial level, only three aspects had a significant effect: Flexibility and Efficiency of Use, Help Users Recognize, Diagnose, and Recover from Errors, and Help and Documentation. The severity rating results showed that most problems were classified as cosmetic, meaning they are not mandatory to fix unless additional time is available. Improvement recommendations are provided to enhance flexibility features, system assistance, and documentation so that Perplexity AI can become more adaptive to user needs. The findings of this study may serve as a reference for the development of AI-based chatbot systems in education and contribute to the academic literature in Human-Computer Interaction (HCI), particularly in the evaluation of digital interface usability.