Uploaded April 2026 | Updated September 2026, 2 days ago
In this episode from the WorkOS booth at HumanX 2026, Michael Grinich sits down with John Wang, co-founder of Assembled, a customer support platform that started in workforce management and evolved into one of the leading AI orchestration layers for support teams.
The counterintuitive finding: AI-powered support doesn't shrink support teams or reduce inbound volume. It increases both. Lower friction means more customers reach out, and businesses end up needing more people to handle the complex cases that AI escalates.
Key topics covered:
• Why AI in support creates induced demand — and why most support teams are growing, not shrinking
• The Jevons paradox of customer service: make it easier to get help, and more people ask for it
• How Assembled decides when to hand off to a human: sentiment, issue type, repeat interactions, and live agent availability
• Why fine-tuned open source models can outperform closed-source models for specific support tasks
• The future of agent-to-agent support — and why AI agents composing support tickets is already happening
• Why AI agents are becoming commoditized, and why the orchestration layer is where the real differentiation lives
Learn more about WorkOS: workos.com and Assembled: assembled.com
In this episode from the WorkOS booth at HumanX 2026, Michael Grinich sits down with John Wang, co-founder of Assembled, a customer support platform that started in workforce management and evolved into one of the leading AI orchestration layers for support teams.
The counterintuitive finding: AI-powered support doesn't shrink support teams or reduce inbound volume. It increases both. Lower friction means more customers reach out, and businesses end up needing more people to handle the complex cases that AI escalates.
Key topics covered:
• Why AI in support creates induced demand — and why most support teams are growing, not shrinking
• The Jevons paradox of customer service: make it easier to get help, and more people ask for it
• How Assembled decides when to hand off to a human: sentiment, issue type, repeat interactions, and live agent availability
• Why fine-tuned open source models can outperform closed-source models for specific support tasks
• The future of agent-to-agent support — and why AI agents composing support tickets is already happening
• Why AI agents are becoming commoditized, and why the orchestration layer is where the real differentiation lives
Learn more about WorkOS: workos.com and Assembled: assembled.com
