In November 2025, Singapore became a focal point for a question many enterprises are now asking: What does meaningful engagement look like in an AI-driven world? Agora brought together industry leaders and innovators for the Conversational AI Mixer: Reimagining Engagement, an evening dedicated to exploring how real-time conversational AI is reshaping the way businesses connect, communicate, and scale.
Partners including Ant Group, OceanBase, Singtel, Wiz.AI, Agora, Akool, Agnes AI, and BetaAI came together to examine how AI-driven conversations are moving beyond automation and into real-world impact across fintech and enterprise engagement.
Held as part of the Convo AI World Event Series, the mixer showcased how conversational AI is reshaping customer journeys, unlocking new revenue opportunities, and redefining how enterprises engage at scale.
Key themes from the evening included:
- Reimagining Enterprise Engagement: How conversational AI is reshaping customer experiences, partnerships, and innovation across banking, insurance, telecom, and Web3.
- Unlocking Fintech Growth: AI-driven engagement models that enable smarter inbound and outbound interactions, while opening new monetization opportunities.
- Experiencing the Future of Finance: Live demonstrations of AI avatars, intelligent hotlines, virtual assistants, and AI-powered customer service in action.
Here’s a summary of each session and what our amazing speakers covered in their presentations:
Session 1: The Future Drivers of Real-Time Conversational AI for Fintech
Sam Yao, Director of Product, Agora
Sam Yao explored the industry’s shift from automation-first AI toward systems designed for understanding, empathy, and meaningful engagement. By combining real-time voice, video, and text with contextual intelligence and deep backend integration, AI dialogue is turning conversation into an API, making natural language the new user interface.
He emphasized that delivering this experience requires a strong ecosystem, capable of supporting real-time infrastructure, intelligence, and end-to-end integration.
Sam highlighted how real-time conversational AI is already reshaping multiple industries:
- Banking & Finance: Instant onboarding, personalized financial guidance, and real-time fraud detection through natural conversations.
- Insurance: Empathy-led claims processing, simplified policy explanations, and proactive risk assessment.
- Telecommunications: Always-on intelligent support with visual troubleshooting and personalized recommendations, without IVRs or long wait times.
- Web3 & Digital Assets: Trust-driven conversations that simplify blockchain concepts, support secure transactions, and enable compliance through clear dialogue.
“We’re moving beyond efficiency toward meaningful engagement, combining the speed of machines with the empathy of humans.”
Session 2: Your Next-Generation AI Database
Davy Feng, Solution Architect - OceanBase International
Davy Feng discussed how the rise of AI-generated content is accelerating the convergence of databases and AI. He emphasized that data is going to stay the foundation of every AI system; from training and validation to continuous iteration, and data quality directly impacts model accuracy and fairness.
He pointed to a major industry shift: vector databases are evolving from standalone tools into native capabilities within general-purpose databases. As AI workloads scale, enterprises increasingly favor unified systems that reduce complexity, lower costs, and support multiple workloads on a single platform.
OceanBase addresses this need by delivering vector search and AI workloads within a cloud-agnostic database, eliminating data silos and operational overhead. With MySQL compatibility and a unified architecture, teams can run transactional, analytical, and AI workloads together.
Key benefits include:
- High-performance vector processing at 10M to 1B+ scale
- Hybrid search across structured and unstructured data using a single SQL query
- Significantly lower storage and memory costs
- Consistent multi-cloud support across AWS, Azure, GCP, and more
“Vector databases are no longer niche tools; they’re becoming a core capability of general-purpose databases.”
Session 3: Step Into the Future: Real-Time Video Interaction with Akool
Shanwei Mu, Head of Algorithms, Akool
Shanwei Mu introduced Akool as a leading AI video platform redefining how creators and enterprises produce and scale visual content. With more than 8 million users and over 300 million pieces of generated content, Akool is powering hyper-realistic avatars, video translation, and personalized media at scale.
At the center of Akool’s platform is its Streaming Avatar, AI-powered digital humans capable of real-time, two-way interaction. With realistic facial expressions, lip-sync accuracy, and emotional nuance, these avatars enable natural video conversations across customer support, training, retail, and live events.
Akool × Agora: Redefining Human–AI Conversations
Shanwei outlined the core challenges of interactive avatars such as latency, emotional realism, expressive voice, and synchronized facial movement, and how Akool addresses them using Agora’s Convo AI capabilities. By combining ultra-low latency delivery with natural conversational flow (including semantic interruption), Akool enables avatars that feel responsive, human, and engaging.
“Akool and Agora are redefining human–AI interaction by making emotionally expressive, real-time video conversations possible at global scale.”
Session 4: Reimagining Revenue Teams Through Human–AI Collaboration
Charles Li, Chief Strategy Office, BetaAI
Charles shared BetaAI’s approach to transforming revenue performance by closely integrating human expertise with AI agents. Originating in digital wealth management, BetaAI today supports over 130,000 bank branches, 1,000+ institutions, and 1.7 million relationship managers, providing deep insight into operating AI at scale in regulated environments.
BetaAI’s model centers on two complementary AI agents. The AI Sales Coach enables teams to simulate real customer conversations, practice objection handling, and build consistency through interactive, two-way simulations, far beyond traditional training methods.
This realism is powered by a layered AI architecture:
- Foundation models enhanced with BetaAI fine-tuning
- Role-based intelligence aligned with industry metrics and guardrails
- AI twins that mirror real customer tone, intent, and objections
The result is a Pro-Coach system that spans the entire sales and service lifecycle, from onboarding to ongoing performance optimization.
“By pairing human judgment with AI readiness and reach, revenue teams can perform better, faster, and at scale.”
Session 5: Bridging the GenAI Gap
Robin Li, Senior AI Strategy & Partnership Director, Wiz AI
Robin Li addressed a growing challenge facing enterprises today: despite $30–40 billion invested in GenAI, nearly 95% of organizations see no measurable ROI. Wiz AI refers to this disconnect as the GenAI Divide, where only a small fraction of AI initiatives reach production and generate real P&L impact.
This gap, Robin noted, is not driven by model quality or regulation, but by how AI is designed, deployed, and embedded into real workflows. While enterprise initiatives stall, employees increasingly turn to personal AI tools that simply work, exposing where current approaches fall short.
To bridge the divide, Wiz AI advocates for AI systems that:
- Learn within enterprise environments
- Adapt continuously over time
- Respect data boundaries and governance
- Integrate seamlessly with existing tools and workflows
Wiz AI supports enterprises across the full AI lifecycle:
- Proof of Concept: Validate ideas with clear, quantifiable outcomes
- Transformation Services: Move pilots into production with governance and change management
- End-to-End Solutions: Deploy AI agents directly into existing enterprise systems
“The GenAI Divide isn’t about better models; it’s about building AI that fits how people actually work.”
Looking Ahead
The Conversational AI Mixer in Singapore highlighted a clear message: the future of enterprise engagement lies in real-time, intelligent, and human-centered conversations. Across fintech, telecom, and enterprise operations, conversational AI is moving beyond experimentation and into real-world impact, powered by strong ecosystems, scalable infrastructure, and thoughtful human-AI collaboration.
The future of conversational AI isn’t about sounding human, it’s about enabling conversations that feel natural, timely, and trustworthy at scale.
Explore how Agora’s Conversational AI Engine provides a complete, real-time stack for building intelligent voice, video, and AI-driven interactions, and see how it fits into your existing architecture through a demo or hands-on evaluation.
