As the prominence of ChatGPT rises, companies are increasingly integrating it into their customer experience initiatives. Artificial intelligence (AI) has become a staple in contact center applications — with AI-powered chatbots offering a streamlined and cost-effective method for delivering real-time customer support. By automating routine tasks and providing immediate access to information, chatbots alleviate the burden on customer service representatives and enhance the overall customer experience.
Beyond basic customer support, chatbots can also serve marketing functions. They can gather data on customer preferences and behaviors, which can then be leveraged to tailor marketing communications and optimize the impact of marketing campaigns.
The adoption of AI in contact centers is accelerating, and companies that hesitate to embrace this technology risk falling behind. Those that effectively implement AI in their customer service strategies will be well-equipped to reap its numerous advantages.
However, chatbots are just one piece of the customer experience puzzle. There are other common use cases of AI in contact centers, including the following:
- Automatic speech recognition
- Agent assist
- Voice/text virtual agents
- Voice biometry
- Voice-to-text and text-to-voice transcription
- AI analytics
- Call/sentiment analysis
- PII/PHI redaction
- Emotional intelligence
- Predictive behavior routing
- Robotic process automation
- Predictive behavioral routing
- Predictive analytics
- Real-time guidance
- Knowledge management
- Workforce automation and training
- To fully appreciate AI’s role in contact centers, it’s essential to understand some fundamental concepts.
Contact Center AI and Generative AI
AI of today differs from pre-AI technology in that AI is learned, whereas traditional systems were programmed. Deep learning, a core technique, has enabled the emergence and further adoption of AI. And while AI has existed for the past five to eight years, a new form of AI technology called GPT-X is challenging older models.
In May 2020, OpenAI, an AI research lab in collaboration with Microsoft, developed GPT-X, a generative AI system capable of mimicking human language. With 175 billion parameters, this deep learning language model was trained on vast text datasets comprising hundreds of billions of words.
GPT-X is a deep learning model that generates structured text sequences based on input text. It excels at tasks such as:
- Question answering: Reduces handle time by offering multifold improvements in agent assist, knowledge base, etc.
- Language translation: Allows for automated in-call translation to all languages.
- Text summarization: Eliminates the need for agent after call work.
GPT-X capabilities include:
- Sophisticated text prediction and auto-generation for responses
- Mimicking language patterns and tone using advance text-to-speech
- Generating creative writing, business memos and functional code, which could include agent follow-up emails, wrap-up codes and other notes
- Producing plausible agent responses to improve handle time and enhance customer satisfaction
However, GPT-X has the following limitations:
- Lack of awareness and abstract reasoning
- Dependence on careful human priming
- Inability to learn from untrained knowledge bases
- Struggles with coherence and meaningful messaging over extended text
- Lack of product, service and business knowledge
- Inability to reason or verify answer correctness
GPT-X vs. Legacy AI Technologies
GPT-X surpasses previous language models by writing its own language and requiring minimal priming. Researchers refer to this as “few-shot learning,” and GPT-X exemplifies its potential. As contact center leaders evaluate GPT-X, it’s important to consider its pros and cons.
- Automation: Reduces time and resources for manual tasks.
- Versatility: Enables agent assistance, call/chat summaries, information extraction and contextual small talk.
- Outdated responses: GPT-X relies on historical data.
- Potential harm: May exhibit bias, harmful language and disinformation.
- Lack of deep world-knowledge: Struggles with laws of physics, math and common sense.
Still, GPT-X presents opportunities for value creation, including bias reduction, measurement, data quality and value-added business strategy. For example, a well-trained AI model can be consistent and unbiased. Whereas human judgement can creep into manual processes, creating bias on external conditions.
GPT-X is revolutionizing AI. Commerce.AI is thrilled to be at the forefront of this transformation with the Commerce.AI auto-MATE™. This product is one way to bring secure, compliant and enterprise-ready generative AI intelligence to the Genesys Cloud CX™ platform. For more information, visit Commerce.AI on the Genesys AppFoundry® Marketplace for the Commerce.AI Base Platform and the Commerce.AI Voice Survey.