Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Recurso digital |
| Lingua: | |
| Pubblicazione: |
Zenodo
2025
|
| Accesso online: | https://doi.org/10.5281/zenodo.15514288 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
Sommario:
- <p> Effective customer service is vital in Nigeria’s competitive power sector, where timely communication drives customer satisfaction. This study designs and evaluates an AI-powered chatbot for Transafam Power Limited to address communication gaps caused by limited traditional channels (e.g., phone calls, emails, WhatsApp) operating only during business hours. Utilizing a sequence-to-sequence deep learning model with attention mechanisms and rule-based natural language processing (NLP), the chatbot delivers 24/7 support, handles routine inquiries, and escalates complex issues to human agents. Trained on a decade of customer interaction data (2014–2023), the system achieves an intent recognition accuracy of 87% and a resolution rate of 85%. Performance is assessed through statistical metrics, including response time, accuracy, and user satisfaction, with significance tested via chi-square analysis. A web-based prototype, developed using HTML, CSS, JavaScript, MySQL, and Python, integrates seamlessly with human agents. Feature importance analysis highlights key predictors of query resolution, enhancing<br>system interpretability. This research underscores the transformative potential of AI-driven chatbots in improving customer service efficiency and accessibility in the power utility sector.<br> Keywords: Customer Service Chatbot, Artificial Intelligence, Natural Language Processing, Transafam Power Limited, Power Sector, Intent Recognition, Statistical Analysis</p>