In today's fast-paced, service-oriented industries, call centers have become indispensable, particularly in sectors that provide critical infrastructure services like natural gas distribution. Traditionally reliant on manual processes and human oversight, these centers face challenges in effectively managing emergency calls. The risk of human error and the limitations of manual evaluation processes compromise the reliability and efficiency of these critical services. This thesis addresses this significant gap by developing and testing an artificial intelligence (AI) model aimed at automating the quality assessment of 187 emergency calls in the natural gas sector. Utilizing a dataset of 187 recorded emergency calls from January to June 2023, this study employs machine learning techniques to create a robust model capable of assessing call quality based on predefined metrics. The findings of this research promise substantial improvements in operational efficiency, the reliability of assessments, and overall emergency management. As such, this study serves as a pioneering effort in integrating AI and machine learning technologies into the emergency call center landscape, offering avenues for future research and practical applications.
Günümüzün hızlı temposu ve hizmet odaklı endüstrilerinde, özellikle doğal gaz dağıtımı gibi kritik altyapı hizmetleri sağlayan sektörlerde çağrı merkezleri vazgeçilmez hale gelmiştir. Geleneksel olarak manuel süreçlere ve insan gözetimine dayalı olan bu merkezler, acil durum çağrılarını etkin bir şekilde yönetme konusunda zorluklar yaşamaktadır. İnsan hatası riski ve manuel değerlendirme süreçlerinin sınırlamaları, bu kritik hizmetlerin güvenilirliği ve verimliliği üzerinde olumsuz etkiler yaratmaktadır. Bu tez, doğal gaz sektöründe 187 acil çağrısının kalitesini otomatik olarak değerlendirmeyi amaçlayan bir yapay zeka (YZ) modelini geliştirerek ve test ederek bu önemli boşluğu ele almaktadır. Ocak ile Haziran 2023 arasında kaydedilmiş 187 acil çağrısının oluşturduğu bir veri kümesini kullanarak, bu çalışma önceden tanımlanmış metrikler temelinde çağrı kalitesini değerlendirebilecek sağlam bir model oluşturmak için makine öğrenimi tekniklerini kullanmaktadır. Bu araştırmanın bulguları, işletme verimliliği, değerlendirme güvenilirliği ve genel acil durum yönetimi konularında önemli iyileştirmeler vaat etmektedir. Bu nedenle, bu çalışma acil çağrı merkezi alanına yapay zeka ve makine öğrenimi teknolojilerini entegre etme konusunda öncü bir çaba olarak hizmet etmekte, gelecekteki araştırmalar ve pratik uygulamalar için yollar sunmaktadır.
In an era of rapid technological advancements and increasing customer expectations, call centers have become critical operational hubs for many businesses. These centers perform multiple functions ranging from customer service to emergency responses. In sectors providing vital infrastructure services such as natural gas distribution, the role of call centers expands to include not just customer satisfaction but also public safety. According to studies by Smith (2020) and Johnson et al. (2019), the effectiveness of these centers directly impacts both consumer trust and the overall safety metrics.
Despite their centrality to both customer service and safety operations, call centers often rely heavily on manual processes and human intervention. In emergency scenarios, particularly for 187 emergency calls related to natural gas distribution, this reliance poses significant limitations and risks. Time-consuming manual evaluation and the vulnerability to human error could lead to delays and mistakes that are unacceptable in emergency situations (Williams, 2018; Davis et al., 2017).
While machine learning and AI technologies hold the potential for automating and optimizing various aspects of call center operations, there is a notable gap in research focused on their application in emergency call handling in the natural gas sector (Brown & Green, 2021). The primary objective of this thesis is to fill this gap by developing and testing an AI model that can automatically assess the quality of 187 emergency calls.
This study will focus on 187 emergency calls recorded between January and June 2023 in the natural gas sector. It will not consider non-emergency calls or calls from other sectors. The limitations include the potential for biases in the recorded call dataset and the challenges related to adapting the findings to different types of call centers.
The potential contributions of this research could revolutionize the call center landscape in the natural gas sector. The automated system aims to improve operational efficiency, reliability, and most importantly, the quick and effective management of emergency calls (Wang & Wu, 2022; Kim et al., 2020; Smith, 2020).