<![CDATA[Abstract: Chatbot performance evaluations are often conducted uniformly without considering operational characteristics across departments. This study aims to evaluate and compare departmental chatbot performance using the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method. The methodology utilizes a secondary dataset of approximately 3,000 interaction logs evaluated against four main criteria: response time, bot confidence score, user feedback rating, and estimated business impact. The evaluation results reveal that Human Resources (HR) achieved the highest rank with a preference score of 0.9531, followed by Marketing (0.6712), Sales (0.4885), Finance (0.3772), and Operations at the lowest rank (0.2356). HR's top performance was driven by the highest satisfaction rating (3.28) and confidence score (0.7986), alongside a fast response time (1,340 ms). Conversely, Operations was hindered by the slowest response time (1,405 ms). In conclusion, the TOPSIS method effectively maps chatbot performance objectively, highlighting that a balance between response speed and answer quality is essential for optimizing user satisfaction. Keywords: chatbot; performance evaluation; mcdm; topsis; departmental comparison Abstrak: Evaluasi performa chatbot sering kali dinilai secara seragam tanpa mempertimbangkan karakteristik operasional antar-departemen. Penelitian ini bertujuan untuk mengevaluasi dan membandingkan performa chatbot berbasis departemen menggunakan metode Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). Metodologi penelitian memanfaatkan dataset sekunder berjumlah ±3.000 log interaksi dengan empat kriteria utama, yaitu response time, bot confidence score, user feedback rating, dan estimated business impact. Hasil analisis menunjukkan bahwa Departemen Human Resource (HR) meraih peringkat tertinggi dengan nilai preferensi 0,9531, disusul oleh Marketing (0,6712), Sales (0,4885), Finance (0,3772), dan Operations pada posisi terendah (0,2356). Keunggulan HR didorong oleh kombinasi rating kepuasan (3,28) dan confidence rate (0,7986) tertinggi, serta waktu respon yang cepat (1.340 ms). Sebaliknya, performa Operations terhambat oleh waktu respon paling lambat (1.405 ms). Disimpulkan bahwa metode TOPSIS efektif dalam memetakan efektivitas chatbot secara objektif, di mana kecepatan respon dan kualitas jawaban terbukti menjadi kunci utama pencapaian kepuasan pengguna. Kata kunci: chatbot; evaluasi performa; mcdm; topsis; perbandingan departemen]]>
