Can AI Teach Empathy and Complex Human Interactions? This Solutionist Says Yes.

Is that wave of the hand inviting or dismissive? Was that tone used warm and reassuring, or curt and condescending?

Human interactions are full of nuance. Much is communicated beyond words, through tone, facial expressions, posture, and gestures. Interpreting them correctly is hard enough. Add cultural differences, and what is friendly in one context may seem rude in another.

That is why role-play has long been the gold standard for building soft skills like negotiation, conflict resolution, and customer service. But traditional role-play training is resource-heavy, difficult to scale, and often intimidating. Digital training solutions have not always bridged the gap as many struggle to create realistic training scenarios or dynamic interactions.

This is where GEM Learning comes in. In partnership with social enterprise Skillseed, GEM Learning has developed an AI-powered platform that makes role-play smarter, more accessible, and scalable. Crucially, it assesses not just voice and content, but also facial expressions, gestures, body language, and tone, where most others solutions focus on voice and content alone. This innovative solution earned GEM Learning and Skillseed a win at IAL’s innovPlus Challenge.

The challenge to develop such a platform was issued by Skillseed, whose core work includes helping clients train their employees in complex human interactions with empathy. The social enterprise’s programmes focus on cultivating dignity, empathy, and cross-cultural engagement and traditionally rely on in-person role-plays, discussions, and facilitator feedback. However, such sessions are time-consuming, limited in reach, and can have varied learning outcomes They are also hard to measure as progress of learners is captured via self-reporting through surveys.

Looking for a more scalable and flexible training modality, Skillseed turned to innovPlus to issue a challenge: develop an experiential learning solution that is intelligently interactive and dynamic, gives real-time feedback to learners, and allows learners to engage on their own terms. Most importantly, the solution needed to support the relational and cultural sensitivity in human interactions within Singapore’s context.

Through innovPlus, Skillseed was matched with EdTech solutionist Gem Learning. Known for their AI-powered training technology, the GEM team had already developed a platform that uses visual intelligence to analyse a user’s tone of voice, facial expressions, body language, and even hand gestures. Feedback is delivered instantly and assesses key metrics such as approachability and empathy.

Before the innovPlus journey, GEM Learning's software was primarily used for behaviour classification tasks like surveillance to pre-empt and prevent potential aggressive behaviour. For the AI to interpret the complex nuances of local communication accurately and meaningfully, more work was put in.





“We had to build datasets based on real interactions in Singapore. This means capturing the diversity of Singapore in both voice and appearance, such as frowns or micro-expressions,” recalled Mr Kamran Kiyani, CEO, GEM Learning. “The variation is wider than we expected! We had to make changes to our AI model to reflect that diversity.”

For example, Mr Kiyani shared, within the Chinese Singaporean community alone, there is a whole spectrum of accents, conversational pace, and preferred facial expressions. The differences are also correlated to factors such as education levels and age. Similar differences existed in Malay, Indian, and Eurasian Singaporeans as well. Creating an accurate, inclusive dataset was a major undertaking.

To deepen the platform’s cultural grounding, GEM Learning also engaged occupational psychologists to interpret behavioural cues and guide the development of realistic, context-sensitive training content. With this foundation, the platform was ready to support Skillseed’s mission of scaling up impactful training.

The advantage of solutioning through innovPlus was the access to IAL’s expertise. IAL coached the team through the process to ensure that the endeavour met key criteria such as scalability and market feasibility.

“IAL made sure that we defined our market and develop a learning solution based on that market’s needs, instead of hyper focusing on Skillseed’s one specific problem or developing a generic solution and hope that someone eventually becomes interested,” shared Mr Kiyani. “That helped us sharpen our focus and we decided on targeting corporate organisations who need to invest in soft skills training for personnel in roles such as customer service.”

The tool does not benefit learners only. It also empowers instructors. The platform allows trainers to use AI prompts to design custom scenarios, tutorials, quizzes, and surveys. They can also define their own evaluation rubrics and track learner performance. “My hope is that this tool will give trainers the opportunity to demonstrate their creativity in coming up with learning solutions!” said Mr Kiyani.

Since its launch, GEM Learning’s platform has been used for a diversity of contexts, from interview training and guest management scenarios in hospitality to sales and customer service training. It has also expanded its platform for international use after adapting its training datasets to function in 14 different cultural contexts, a testament to its flexibility and scalability. The company has also secured its first overseas contract customer in Turkey.

Soft skills are more vital than ever in the fast-evolving global economy. GEM Learning’s AI-powered role-play training reimagines experiential learning in the digital age and helps to unlock human potential, one AI-simulated scenario at a time.