Empathetic AI: How GenAI Virtual Agents Will Be Leveraged

Imagine you're shopping online or in a store and have a question about a product. But you don’t want to find a store employee or call into their contact center to speak with a salesperson while you wait for answers.

With a simple snapshot of a QR/UPC code on the product or a tap on your mobile device within an application, you're instantly connected to a virtual agent via video chat. This AI assistant not only understands your questions but can also pick up on your facial expressions and body language via your video discussion, to respond with empathy and some level of emotional intelligence.

"I can see you're a bit frustrated," the virtual agent says with a warm, reassuring tone. "Let me walk you through the product details and address any concerns you might have." As you interact, the AI seamlessly adjusts its language and mannerisms, mirroring your communication style to create a more engaging experience. But this virtual agent isn't just friendly – it's a veritable product expert, able to provide in-depth information, answer follow-up questions, and even offer recommendations based on your preferences and purchase history in that store or with that brand.

"Given your interest in sustainable fashion, you might also like our new line of eco-friendly accessories," the AI suggests, presenting visuals of complementary items tailored to your taste or directing you to a different area in the store. This level of personalized, emotionally intelligent, and visually present interaction is poised to become the norm as generative AI virtual agents continue to advance rapidly on our devices. Powered by large language models, computer vision, personalization data and specialized emotionally centric AI models, these virtual assistants will bridge the gap between digital and human experiences.  

In the realm of customer support, generative AI virtual agents will be able to engage in natural, multi-intent driven conversations, understanding context and providing relevant, empathetic responses. They'll leverage emotion recognition APIs to detect and respond appropriately to a customer's emotional state using your voice tones and facial reactions, offering reassurance or escalating to a human agent as needed.  While using this same API info to mimic caring facial expressions that match the sentiment of the avatar it’s using to support you.

For sales and marketing, these AI assistants will become powerful tools for personalized recommendations, product demos, and even sales support. By analyzing a customer's behavior, preferences, and real-time reactions, gen AI agents will tailor their approach, highlighting relevant features and addressing concerns in a way that resonates on a personal level. 

With a specific voice/accent, appearance or quirk of your choosing potentially.  If the app or brand allowed it, I’d choose a humorous avatar, who isn’t afraid to give me a tough time or poke fun at me.  Snark, with a side of sarcasm would be my prompt request if I am dealing with a human or AI.  With AI, you will eventually get to choose, beyond just the male, female or accent type used. What would you choose? 

As these capabilities evolve, customers may prefer interacting with a custom prompted AI assistant over a human agent for specific needs. Unless we start measuring our human agents’ conversational styles/personas and intelligently routing our contacts based on the chosen customer preference.

For straightforward inquiries or transactions, the convenience, personalization, and emotional intelligence offered by gen AI virtual agents could make them the preferred choice for particular use cases, especially for customers who value efficiency and a seamless digital experience. If the AI is trained well enough.

Unlike traditional rule-based or machine learning-based chatbots, gen AI virtual agents can engage in more natural and contextual conversations, understanding nuances and providing more relevant responses in most cases.

The evolution of chatbots, while useful for handling simple, structured inquiries, often struggled with complex or open-ended queries, leading to frustrating experiences for customers. Gen AI virtual agents, however, can understand and respond to intricate requests. They are already scheduling appointments, taking after-hours calls and providing troubleshooting or product support. 

Now add in the 2-3 other reasons you’ve been meaning to call in.  Maybe you won’t need to be transferred elsewhere to do everything you need within the original call to AI.

In the healthcare sector, gen AI virtual agents are being used for triage and initial symptom assessment, providing patients with personalized advice and guidance while alleviating the burden on human medical staff or nurses for specific early-stage care cases. Additionally, counseling and mental health services are exploring the use of gen AI virtual agents as an initial point of contact, offering empathetic and non-judgmental support to individuals in need.

In situations where requiring a human is not dire or critical.  Entire groups are being formed to study, measure and create guidelines to identify whether AI can truly become emotionally intelligent, according to Emotional Intelligence Author and expert, Daniel Goleman. 

Artificial general intelligence (AGI) developments and other Gen AI innovations have had many contact center as a service companies scrambling to create partnerships to ensure their tools can stay up-to-date with the most relevant features of the future.  Whether you’re planning for this now or intend to closely watch the large enterprise developments in 2024, there are a few considerations that we’re discussing with ICMI training and consulting clients now: 

Benefits of Generative AI Virtual Agents

• Improved Customer Experience: Gen AI virtual agents can engage in more natural and contextual conversations than yesterday’s chatbots, leading to better customer satisfaction and loyalty for users who prefer to use gen AI. 

Challenges and Considerations

While the potential benefits of gen AI virtual agents are significant, there are also important challenges and considerations to address:

• Data Privacy and Security: As gen AI virtual agents handle customer data and interactions, it is crucial to ensure that proper data privacy and security measures are in place to protect sensitive information.  Data readiness obviously is required as well.
• Bias and Fairness: Gen AI models can potentially inherit biases present in their training data, leading to unfair or discriminatory outputs. It is essential to train these models on diverse and unbiased datasets and continuously monitor their performance for potential biases.
• Human Oversight and Integration: While gen AI virtual agents can handle many interactions independently, human oversight and the ability to escalate to human agents for complex or sensitive issues is still necessary. Seamless integration between AI and human agents is vital for optimal customer or employee experiences.

To successfully implement gen AI virtual agents in contact centers or service desks, organizations should follow best practices and strategic approaches:

• Start with Low-Effort Use Cases: For organizations new to gen AI, it is advisable to begin with low-effort use cases, such as FAQ assistance or basic inquiries, to gain familiarity and build confidence in the technology.
• Moderate Use Cases: As organizations become more comfortable, they can explore moderate use cases like personalized product recommendations, appointment scheduling, or initial triage in healthcare settings.
• Complex Use Cases: For more mature organizations, complex use cases such as in-depth sales qualifications, customer onboarding, or higher tiered technical support can be tackled with gen AI virtual agents, leveraging their ability to engage in multi-intent, contextual conversations.
• Training and Fine-Tuning: Effective training and fine-tuning of gen AI models on domain or company specific data and use cases are crucial for optimal performance. Continuous monitoring and refinement based on user feedback are also essential.Change Management and User Adoption: Implementing gen AI virtual agents requires careful change management strategies to ensure user adoption and seamless integration with existing processes and human agents.  For many, it is too early for this effort.

The future of gen AI virtual agents in contact centers is promising however, with ongoing advancements in natural language understanding, multimodal interactions (combining text, speech, and visuals), and personalized conversational experiences.

However, ethical considerations and responsible development of these technologies must remain a top priority. There will be many customers who still prefer a human experience in all interactions.  But, just like the slow uptick of live chat usage externally and instant messaging internally, the Gen AI chatbot or virtual agent is already a force for those organizations who have built it with their users in mind. 

Someday, you'll be able to video chat with a virtual AI agent that does these things. Just don't be surprised if it occasionally roasts your questionable fashion choices if that is the prompt or avatar you have chosen. Who knows, you might just make a new AI friend in the process or have an AI celebrity avatar to help you out.