A CNN-Based Model for Intelligent Chatbots in Digital Marketing for Banking Services

The model that this paper proposes to use to create smart chatbots that can be employed in banking services to achieve the digital marketing is based on Convolutional Neural Network. To facet customer queries and prepare rich word embeddings that capture contextual information, the model adopts sophisticated natural language processing techniques such as using the SpaCy to clean text and tokenize customer query so that it can be prepared using Fast Text API that provides rich word embeddings. The trained CNN architecture that was constructed with the assistance of the TensorFlow with the Keras API becomes efficient at classifying intents of its users, which can be used to answer them specifically and with relevance. Experiments have shown that the model has a high accuracy reach of 92.3 percent and low average response time of 0.6 seconds; higher than the various conventional models such as LSTM and GRU. Moreover, it has been included in their digital marketing systems providing better measures of customer interaction, delivering to the point of a 17 percent growth of campaign click through rates and a 12 percent improvement in the conversion of leads. This study points to the possible capabilities of the CNN-based chatbots in boosting the efficiency of operations and customer engagement within the scope of digital marketing activities in the banking industry