Conference Presentation, Fireside Chat, Interview
Adrian McDermott, CTO, Zendesk: What CX Teaches Us About AI And The Future Of Work
- Zendesk projects the global customer service economy will shift by the end of the decade from a $500 billion human-driven and $100 billion software sector to approximately $250 billion in software and $100 billion in humans, anticipating that 51% of customers currently prefer resolving issues with bots and this proportion will grow year over year.
- By 2030, the landscape is expected to transition from human-centric operations to agent-powered interactions for the majority of cases, with AI capabilities extending beyond customer service to broader industries and businesses.
- Operational efficiency is predicted to improve with AI reducing time to first response to approximately 200 milliseconds, while metrics like average handle time and ticket volume may become less relevant; conversely, the "Jevon's paradox" may increase total service volume due to greater convenience, and automation is expected to drive increased engagement on phone channels when automated paths fail.
- Zendesk plans to launch voice beta functionality in the third quarter using speech-to-speech technology, introduce a pricing system separating human and automated resolutions to potentially allow enterprise-wide licensing, and experiment with outcome-based pricing models as SaaS globally shifts toward metered usage tied to delivered value.
- The strategy involves empowering "citizen developer" agents to build integrations and procedures, freeing teams to focus on high-value interactions like churn prevention and upselling, while utilizing insights from the 71% of tickets representing business failures to fix underlying product issues directly.
- Zendesk anticipates customers will increasingly view support as a revenue source rather than a cost center, citing data that 54% of contacts drive 95% of growth, though some risk-averse clients will require vendor assistance in identifying intents and building MCP-based AI actions.
- Historical precedents from the 2010 introduction of robots at Amazon and the banking shift following ATMs suggest AI will expand workforces and transform roles rather than eliminate jobs, with a broader industry trend of re-hiring humans for complex, high-value VIP support.
- Significant risks include the possibility of moving too far ahead of customer readiness, potential client attrition due to perceived vendor misunderstanding, and customer fears regarding GDPR compliance that must be addressed even if not perceived as real by the vendor.
- The net result is expected to be more winners than losers for society, although some individuals may feel aggrieved or displaced during the transition, and psychological preferences for machine interaction are driven by a desire to avoid the emotional burden of "making a person suffer" with a problem.
- Product teams may leverage AI-driven productivity to build more features, potentially generating more failures and support tickets in a self-perpetuating cycle, while the distinction between fixing tickets via agents and fixing products remains a key strategic differentiator.