More Human Than Human

Happy Monday!

Last week, a study published in Nature Communications Psychology dropped a bombshell that should make every customer service executive stop and reconsider what "human touch" really means. Six AI models, including ChatGPT-4, outperformed humans on five standardized emotional intelligence tests with an average score of 81% versus just 56% for humans. The AI systems demonstrated superior ability to suggest appropriate responses in emotionally charged situations and understand the full context of the conversatio.

The Empathy Paradox

We're witnessing a fascinating paradox: machines that don't feel are becoming better at emotional intelligence than the beings who created them. While humans invented empathy, we're surprisingly bad at consistently applying it. This is especially true in high-stress, high-volume customer service environments.

The reason isn't that AI has achieved consciousness. It's that AI has solved a fundamental problem humans struggle with: the difference between listening to understand and listening to respond.

The Meta Trend: Context Extraction vs. Reactive Processing

My hypothesis about why AI is excelling where humans falter is that humans often listen to respond and are reactive to what is being said, but AI extracts full context from every interaction before responding.

When a frustrated customer calls, human agents are typically formulating their response while the customer is still talking. They anticipate objections, prepare defensive explanations, or mentally rehearse standard scripts. This reactive processing means we miss nuanced context, underlying concerns, and emotional subtleties.

AI systems, by contrast, process the entire conversation before generating responses. They analyze sentiment, extract entities, maintain conversation history, and consider multiple contextual factors simultaneously. They don't get defensive, tired, or emotionally triggered.

Pattern Recognition: Three Ways AI Is Becoming More Human

Pattern #1: Superior Emotional Intelligence in Real-World Applications

The Nature study isn't just academic theory. These findings are being validated in practice. Research shows that AI can provide "higher quality yet less empathy" in customer service, but customers increasingly prefer this consistent competence over inconsistent human empathy.

Consider this: Almost one-half of customers think AI agents can be empathetic when addressing concerns, and 69% of organizations believe generative AI can help humanize digital interactions. Customers don't need agents to feel their emotions, but they do need agents to understand and appropriately respond to them.

Pattern #2: Context Retention That Surpasses Human Memory

Modern AI-powered chatbots ensure that all relevant context, including the customer's query and interaction history, is transferred during handoffs. This minimizes repetitive explanations and creates a much smoother experience. Humans, meanwhile, routinely ask customers to "repeat their issue" or fail to review previous interactions.

The difference is architectural: AI systems maintain conversation history and relevant contextual information to ensure responses are appropriate to the ongoing dialogue rather than treating each interaction in isolation. This context awareness allows AI to provide more coherent, personalized service than human agents juggling multiple conversations while dealing with system limitations and time pressure.

Pattern #3: Consistent Performance Without Emotional Fatigue

While humans may not always exhibit maximal performance in emotionally charged situations due to factors like mood, fatigue, personal preferences, or competing demands, AI systems like ChatGPT-4 can reliably deliver maximal performance in emotional understanding and management in every interaction.

The data backs this up: Organizations operating or optimizing AI-powered customer service reported 17% higher customer satisfaction, while 58% of CX leaders believe their chatbots will grow more advanced in 2024, and 68% of consumers believe chatbots should have the same level of expertise and quality as highly skilled human agents.

The Contrarian Take: Redefining "Human Touch" in Customer Service

The conventional wisdom says customer service needs more "human touch." But what if we're defining human touch incorrectly?

Most customer service interactions don't require deep emotional connection. They do, however, require competent problem-solving, accurate information, and appropriate emotional responses. AI is proving it can deliver all three more consistently than humans.

The real insight: We've conflated the ability to have emotions with the ability to handle them appropriately. The Nature study shows AI can suggest better emotional responses precisely because it isn't clouded by its own emotional reactions.

Consider the irony: After claiming its AI chatbot could do the work of 700 representatives, Klarna is turning back to people to handle more of its customer service work. Despite the narrative, this is not because AI performed poorly. Their AI still handles two-thirds of inquiries with 82% improved response times and 25% fewer repeat issues. They're adding humans back for the small percentage of cases requiring genuine emotional connection.

This suggests the future isn't AI vs. humans, but a combination of AI and humans each playing to their strengths. AI can handle the majority of interactions that require emotional intelligence but not emotional connection, while humans can focus on the truly complex cases that need empathy, creativity, and judgment.

Practical Implications: What This Means for CX Leaders

For Customer Experience Teams: Stop measuring AI success against human empathy and start measuring it against human emotional intelligence. AI can simulate certain conversational responses or track behavioral trends, but it cannot authentically connect. However, many AI applications may not require this to achieve their intended outcomes.

For Training and Development: CX training often focuses on understanding AI responses, managing customer interactions, and handling case escalations, ensuring CX teams are well-prepared to leverage AI effectively. The new skill isn't competing with AI on emotional intelligence, but knowing when human intervention adds value beyond what AI can provide.

For Technology Investment: An impressive 65% of respondents intend to expand their use of artificial intelligence (AI) in customer experience (CX) over the next 12 months. But focus on context-aware AI systems, not just fast responders. AI that considers prior interactions ensures coherent, context-aware responses and crafts responses based on input and context, addressing user queries effectively.

The Revenue Model Shift: Gartner predicted that by 2025, 80% of customer interactions will involve AI, but the most successful implementations combine AI's consistent emotional intelligence with strategic human escalation. Companies achieving ROI of over 1000 EUR from single campaigns are those using AI to handle emotional intelligence at scale while reserving human agents for complex, creative problem-solving.

AI has massive potential to drive real value in support and CX use cases. Thinking of it as an AI or human problem is not the right solution. Leveraging AI’s ability to perceive context and address emotionally charged conversations with a clear head can help deflect or deescalate support interactions before they become larger issues. In this case, AI acts as a front-line defense to aid and support the humans behind the line who are waiting to do what humans do best: create deep emotional connections with customers and solve problems creatively.

In motion,
Justin Wright

1. Record Labels in Talks to License Music to AI Firms (Bloomberg)

2. Meta Aims to Fully Automate Ad Creation Using AI (Wall Street Journal)

3. Introducing Bing Video Creator (Microsoft)

4. AI that improves itself by rewriting its own code (Sakana)

5. Introducing AI Studio Avatars (HeyGen)