Table 1 Overview of conversational AI... | Emerald Publishing

Table 1

Overview of conversational AI applications in text-based customer service

	Conversational analytics	Conversational coaching	Chatbots
Description of AI application	Conversational analytics extracts data from conversations using natural language processing	Conversational coaching teaches conversation partners to use verbal and non-verbal strategies to improve communicative interactions based on conversational data with natural language processing	Chatbots conduct conversations via text automatically with their users
Role in customer service	Gather insights from previous and ongoing customer service conversations and feedback these to customer service agents	Recommend actions to customer service agents based on gathered insights from conversational data	Customer service chatbots replace human customer service agents and automatically respond to customer inquiries
Type of service encounter	AI-supported service encounter The customer directly interacts with a human (customer service agent), who is supported by AI	AI-supported service encounter The customer directly interacts with a human (customer service agent), who is supported by AI	AI-performed service encounter The customer directly interacts with an AI
Customer service tasks	• Understand customer journey • Connect touchpoints (live chat, SMS, social media) • Enforce compliance in call centers • Root-cause analysis in conversations	• Auto-completing messages • Recommending adjusting tone or volume (in voice) • Transferring customers to other agents • Recommending discounts • Recommending appropriate pre-written responses	• Automate FAQs • Gather initial information • Track orders and deliveries • Appointment scheduling • Other company-specific customer service tasks
Common AI techniques	Natural language understanding techniques • Sentiment analysis • Part-of-speech tagging • Intent detection • Summarization • Topic modeling	Natural language understanding techniques as in conversational analytics to gather insights plus • Recommender systems • Reinforcement learning	Chatbots also rely on natural language understanding and decision-making techniques but also require • Natural language generation (e.g. language models) • Dialogue state tracking
Opportunity for relational personalization	Particularly suited to extract knowledge about customer • Customer journey mapping • Sentiment analysis • Emotion analysis • Customer profiling	Particularly suitable for connecting conversation styles with customer knowledge • Recommend personalized level of conversation style (e.g. less repetitive) • Suggest use of emojis based on customer emotion • Route customers to appropriate agent based on personality • Suggest more appropriate (e.g. empathetic) response	Chatbots combine the other two applications and are particularly suitable for realizing personalized conversation styles • Generating personalized responses • Adjusting styles throughout conversations

	Conversational analytics	Conversational coaching	Chatbots
Description of AI application	Conversational analytics extracts data from conversations using natural language processing	Conversational coaching teaches conversation partners to use verbal and non-verbal strategies to improve communicative interactions based on conversational data with natural language processing	Chatbots conduct conversations via text automatically with their users
Role in customer service	Gather insights from previous and ongoing customer service conversations and feedback these to customer service agents	Recommend actions to customer service agents based on gathered insights from conversational data	Customer service chatbots replace human customer service agents and automatically respond to customer inquiries
Type of service encounter	AI-supported service encounterThe customer directly interacts with a human (customer service agent), who is supported by AI	AI-supported service encounterThe customer directly interacts with a human (customer service agent), who is supported by AI	AI-performed service encounterThe customer directly interacts with an AI
Customer service tasks	• Understand customer journey • Connect touchpoints (live chat, SMS, social media) • Enforce compliance in call centers • Root-cause analysis in conversations	• Auto-completing messages • Recommending adjusting tone or volume (in voice) • Transferring customers to other agents • Recommending discounts • Recommending appropriate pre-written responses	• Automate FAQs • Gather initial information • Track orders and deliveries • Appointment scheduling • Other company-specific customer service tasks
Common AI techniques	Natural language understanding techniques• Sentiment analysis • Part-of-speech tagging • Intent detection • Summarization • Topic modeling	Natural language understanding techniques as in conversational analytics to gather insights plus• Recommender systems • Reinforcement learning	Chatbots also rely on natural language understanding and decision-making techniques but also require• Natural language generation (e.g. language models) • Dialogue state tracking
Opportunity for relational personalization	Particularly suited to extract knowledge about customer• Customer journey mapping • Sentiment analysis • Emotion analysis • Customer profiling	Particularly suitable for connecting conversation styles with customer knowledge• Recommend personalized level of conversation style (e.g. less repetitive) • Suggest use of emojis based on customer emotion • Route customers to appropriate agent based on personality • Suggest more appropriate (e.g. empathetic) response	Chatbots combine the other two applications and are particularly suitable for realizing personalized conversation styles• Generating personalized responses • Adjusting styles throughout conversations