TL;DR:
- Generative AI is making its mark in contact centers, with various use cases emerging.
- The current focus is on assisting agents rather than customer-facing scenarios.
- Abstract summarization is considered a highly valuable application of generative AI.
- Contact centers can generate call transcripts and use generative AI to summarize key information.
- Generative AI excels at summarization, saving significant time for agents and reducing costs.
- Five9’s offering already includes abstract summarization, with other features in development.
- Intent classification can potentially shift from NLP models to generative AI.
- Generative AI can simplify entity extraction tasks, such as breaking down spoken addresses.
- It enables contact centers to extract valuable business insights from conversational data.
- Generative AI aims to free up agents’ time, allowing them to focus on meaningful conversations.
- Experts emphasize the importance of considering the agent’s experience in implementing generative AI.
- The potential for generative AI to enhance the agent’s role and improve customer service is significant.
- Decision-makers should resist perceiving generative AI solely as a cost-cutting, customer self-service tool.
Main AI News:
The integration of generative AI in contact centers is rapidly gaining traction, revolutionizing the industry and unveiling a plethora of use cases. Although the current applications predominantly focus on assisting agents rather than customer-facing scenarios, the allure of agent replacement looms tantalizingly close, enticing contact center decision-makers to explore this transformative technology further.
One such use case that has emerged as a clear winner is an abstract summarization. Esteemed analyst Dave Michels of TalkingPointz, during a recent No Jitter webinar sponsored by Five9, declared summarization as the “killer app” of the moment. Demonstrating its functionality, Richard Dumas of Five9 explained how the contact center system generates a transcript of a customer call, which is then presented to GPT-3, the esteemed large language model preceding the latest GPT-4, responsible for powering the renowned ChatGPT chatbot.
In this specific use case, GPT-3 is tasked with summarizing the call and extracting key information provided by the agent, such as the customer’s name, address, and mentioned products. The agent has the option to review and modify the summary as needed. Michels praised the exceptional summarization capabilities of the language model, emphasizing its efficiency and speed in capturing the salient points of the conversation. Leveraging generative AI for agent wrap-up activities in contact centers translates to significant time savings. Even a minute saved from a five-minute call equates to a remarkable 20% cost reduction for the contact center, as highlighted by Dumas.
The abstract summarization feature is already available in Five9’s comprehensive offering. Furthermore, several other promising features are currently in beta testing or nearing release, including:
1. Intent Classification: Traditionally reliant on natural language processing (NLP) engines in conversational AI for intent classification, contact centers are now exploring the possibility of harnessing generative AI for this task. Dumas revealed that preliminary experiments indicate that large language models excel as classifiers. GPT-3 presents an alternative to NLP models for intent detection and classification.
2. Entity Extraction: Generative AI proves invaluable in scenarios such as extracting spoken addresses from callers. Instead of the sequential querying of the street address, city, state, and zip code, the language model-powered system can request the caller to provide the complete address, subsequently employing generative AI to extract and segregate the individual elements automatically.
3. Insights: Generative AI facilitates expedited and simplified extraction of valuable business insights from conversational data. For instance, it enables the analysis of call reasons (e.g., exchange or refund requests) in conjunction with metrics like average handle time, average hold time, and average queue time.
Michels and Dumas emphasize that these use cases primarily aim to free up agents’ time, allowing them to engage in more empathetic and responsive conversations with callers. By offloading tasks such as transcription, note-taking, and eventually dispositioning to AI-powered systems running in the background, agents can focus solely on fostering meaningful interactions with customers.
The focus on agent-assist functions for generative AI resonates profoundly with contact center expert Amas Tenumah, author of the acclaimed book, “Waiting for Service: An Insider’s Account of Why Customer Service is Broken + Tips to Avoid Bad Service.” In a recent conversation, Tenumah voiced concerns that contact centers might adopt generative AI primarily as a means to provide faster and cheaper customer service, overlooking its true potential in enhancing the agent experience.
He astutely observes that the most critical human element in contact centers, the agent, has often been neglected. Cognitive overload plagues agents, and generative AI possesses the power to alleviate this burden. Tenumah envisions a generative AI system that actively listens to callers and instantly equips agents with relevant supporting information, empowering them with a unique “superpower.”
“If we direct this tool toward the right individuals, rather than focusing solely on consumer-facing aspects, I firmly believe it will not only satisfy our customers but also deliver superior returns on investment for our shareholders,” Tenumah affirms.
Conlcusion:
The integration of generative AI in contact centers signifies a transformative shift in the market. The current focus on assisting agents and improving their efficiency through applications like abstract summarization and intent classification showcases the potential of generative AI in enhancing agent-customer interaction. By automating tasks and providing real-time support, generative AI enables agents to focus on meaningful conversations and deliver more personalized and empathetic customer service.
This not only improves the overall customer experience but also leads to significant cost savings for contact centers. As the market continues to explore the capabilities of generative AI, decision-makers must recognize its potential to revolutionize the agent experience and drive positive outcomes for both customers and shareholders. Embracing generative AI as a tool to empower agents and optimize customer service will position organizations at the forefront of the evolving contact center landscape.
