Balancing automation and empathy: How AI can deliver empathy at scale

As artificial intelligence (AI) continues to reshape the landscape of consumer engagement, many consumers are concerned about the loss of human touch in AI. According to a recent YouGov survey, nearly half (48%) of consumers in Hong Kong, Indonesia, and Singapore express concern over the diminishing human element in AI-driven experiences. This increases to 50% among Singaporeans.

Interestingly, generational attitudes vary widely. Gen Z appears more comfortable with AI's increasing role, with only 43% expressing concern - compared to a significant 61% of Baby Boomers who feel uneasy about the shift. Meanwhile, Indonesians show the least concern overall, suggesting cultural and regional differences in how AI is perceived.

Don't miss: Survey: 48% of HK, SG and ID consumers concerned about loss of human touch in AI

Automation vs empathy

Whenever a new technology comes along, it will create tension between the old familiar ways and the new ways that seem less emotional and more clinical, according to Sandeep Joseph, co-founder and CEO, Ampersand Advisory. "We have seen this at every inflection point of technology, and we saw it with the rise of the internet too. You could argue that community management at scale, or loyalty and CRM have also become impersonal, as technologies to deliver consumer connect have advanced."

While some are worried about the lack of empathy in AI, Joseph believes AI is improving by leaps and bounds every day and will soon be able to simulate empathy and emotions well. "In recent research, consumers could not tell if a response came from AI or a human. Even in 2023, ChatGPT could provide responses that ranked a lot higher than human responses, when it came to empathy. So I just see AI continuously getting better and more emotional," he added.

The growing tension between technological advancement and emotional connection, reflects a growing awareness that while automation delivers efficiency, it can compromise emotional connection - especially in service industries where empathy and human nuance are essential, according to Leo Siu, director at Cloudbreakr. He added:

Consumers, particularly older demographics and those in high-service-expectation markets such as Singapore, are cautious about technology replacing personal interaction rather than enhancing it.

The tension between automation and empathy arises mainly from concerns that poorly configured AI systems can produce subpar answers, leading to unmet customer needs, said Joy Liu, head of marketing, SleekFlow.

“In SleekFlow’s research, we actually found that in SEA consumers that welcome AI in brand communication, the top three rated traits for AI agents were: Fast checkout or booking, response speed, and multilingual support; while the top three traits for human agents were having empathy and emotional intelligence, the ability to solve complex problems and personalisation,” she added.

Meanwhile, Jacky Chan, CTO, Votee AI sees the current development of AI as heading into the "uncanny valley", a term coined by Japanese robotics professor Masahiro Mori in 1970. “When the resemblance becomes very close but not quite perfect, our positive feelings can plummet, creating a ‘valley’ in a graph that plots affinity against human likeness.”

He noted that while some of the latest LLMs, such as GPT, might appear empathetic, there's a debate on whether AI can genuinely possess empathy or only simulate it, as it lacks a "heart." He points to the rise of companion AI and virtual girlfriends as a controversial area highlighting this tension.

Existing emotional intelligence

In light of the growing demand for human touch in AI, several key technologies are being adopted to make AI more emotionally aware, said Cloudbreakr’s Siu. “This includes sentiment analysis which detects user emotions through text or voice tone; long-term memory in conversational agents such as ChatGPT that enables continuity and personalisation.”

Other examples include retrieval-augmented generation (RAG) which grounds AI responses in real-time, relevant data, as well as multimodal AI that integrates voice, vision, and text to better interpret user intent and context.

Dominique Rose Van-Winther, chief AI evangelist, Final Upgrade, said today’s emotional AI is technically dazzling. "It reads your facial expressions, vocal tone, syntax, even your typing rhythm. Some of the top tools and systems include affective computing platforms such as Affectiva that decode micro-expressions; multimodal emotion recognition integrating voice, visual, text, and physiological signals; as well as cross-cultural sentiment models that attempt to map emotional nuance in different languages."

Centred around that human touch is "artificial empathy", said Votee’s Chan. This involves enabling AI to detect human emotions through various cues such as facial expressions, body gestures, speech descriptors and sentiment analysis, he added.

He added that maintaining a human touch is achievable with current AI, especially in consumer engagement. “I believe chatbots are already capable of handling customer support and self-help interactions effectively.”

A successful example he provided is UBTECH's Alpha Mini Robot, which utilises Smart Robot's “Maatje” software. She added:

These robots are being prepared to offer healthcare services to children at Sint Maartenskliniek in the Netherlands, having been trained in exercise, empathy, and cognition capabilities.

Can AI be more humanised?

On the tech front, AI can retain a human touch by combining automation with thoughtful design including hybrid models, said Cloudbreakr’s Siu. “AI can be used for routine queries, escalating emotional or complex issues to human agents. It can adjust messaging based on detected sentiment; as well as leveraging memory and context to make interactions feel familiar and relevant.”

“AI should be developed not just for functional accuracy, but with a focus on human-centred design. Transparency, cultural sensitivity, and user control over personalisation and memory are critical for long-term trust and adoption - especially in diverse, high-expectation markets,” he added.

Another solution is combining the strengths of automation with empathetic human interaction, creating a seamless and personalised customer experience, said SleekFlow’s Liu. “AI Agents can provide real-time, personalised responses to routine queries, drastically reducing waiting times and ensuring continuous support without overwhelming human agents.”

This allows human representatives to focus on complex, empathy-driven conversations that require active listening and emotional intelligence, she added. “SleekFlow’s agent flow could be configured to understand when an AI conversation is best handed off to a human agent to follow up, allowing businesses to balance the need to automate routine tasks and taking care of customer’s emotional needs.”

Attaching emotional language to machine outputs without clear boundaries creates manipulation, said Van-Winther, adding that users should not be nudged into emotional connection with a system that cannot reciprocate genuine care. "AI can be present, responsive, and non-judgmental. These are powerful advantages. We should focus on these instead of manufacturing artificial warmth."

"Let humans handle complexity and nuance. Let AI do what humans can't: scale, speed, pattern recognition. Embrace transparent artificiality. Say what you are. Own your boundaries. Earn trust through consistency, not mimicry," she added.

To maintain a human touch, the technology should be monitored, and real humans need to sprinkle in human touches, said Joseph. "Quirkiness, humour, empathy and generally resonating with the emotional state of the consumer are ways that brand managers and agencies can improve their own EQ and leverage AI to improve its perceived emotional connection." He added:

The secret of successful AI application for CRM and loyalty will lie in the effort the brand puts in to differentiate itself while still being relevant and helpful.

Join us this coming 17 June for #Content360 Hong Kong, an insightful one-day event centered around responsible AI, creativity VS influencers, Xiaohongshu and more. Let's dive into the art of curating content with creativity, critical thinking and confidence!

Related articles:

Agentic AI and the CX reset: The tech changing how brands serve and retain customers
Agentic AI for dummies: 101 on how marketers can leverage on the trend