When someone contacts customer service or a helpdesk, their goal is not to have their call logged. Their goal is to have their problem solved. Whether that's a technical issue with an application or negotiating a refund or product return, customer service is about outcomes and not tickets.
Yet the productivity of support teams is typically measured by number of calls taken and closed. Whether those closed calls have delivered a successful outcome is a different metric that is not always measured and can have a dramatic impact on revenues. According to Zendesk's recent CX Trends Report, 89% of CX leaders say one unresolved issue is enough to lose a customer.
The good news is that customers are becoming more accustomed to having AI assist them through their journeys with brands. More than half of consumers report they are comfortable relying on personal AI assistants for everyday tasks. As Zendesk's APAC CTO David Hall-Johnston explained, this is a major driver behind the company's new Outcome AI Platform.
"AI matters when it improves people's lives in tangible ways. It's not just about productivity," he said.
"We see that as an opportunity to move beyond delivering service by closing tickets, to helping people achieve the outcome they came for. The platform connects intent to action and action to measurable outcomes."
The Outcome AI Platform integrates Open Service Directory, Zendesk Bridge, Specialised AI Agents and Self-Improving Zendesk. Together, they leverage AI agents, integrations with other platforms and information sitting in knowledgebases to find and solve customer queries and issues as quickly as possible. When an AI successfully resolves an issue, the processes it followed are added to Self-Improving Zendesk to ensure the same issue can be resolved in future.
This is important. When AI solves a new problem the first time it uses probabilistic methods. Once the problem is understood, it uses deterministic processing which is more efficient.
Specialised AI agents are tailored to the needs of specific industry verticals. Commerce is the first set of agents Zendesk is bringing to the platform, with plans to add financial services, media and technology.
While AI can be a powerful way to enhance customer experience, it is important to retain people in the process. As Hall-Johnston explained, working out where to involve people in processes requires careful consideration.
"Our customers decide where to draw that line. There are many tasks that can be fully automated, like simple question and answer queries, but we understand there are highly complex and stressful tasks where human empathy is required or where brands need human judgement. The Outcome AI platform delivers all these outcomes. That could be a human at the front-end interfacing via voice or chat. Or it could be a customer chatting with an AI chatbot but then a human in the background decides whether to approve a request," he said.
There has been concern that introducing automation into customer service functions could lead to job losses. But Hall-Johnston says Zendesk's research and his observations of what is happening in industry suggests the opposite is true. According to Zendesk data, among teams with a median automation rate of 41%, total customer interactions roughly doubled.
"This means AI is absorbing demand businesses never could have afforded to serve with people alone."
"What we see is good customer service creates more requests for customer service. So, the volumes in some interactions are increasing. Adrian McDermott, our global CTO, calls this the 'service debt'. We are helping our customers pay that down by recognising that there are unmet requests for service. People may have dropped off hold or had a bad experience with a chatbot. What we're seeing now is good service breeds more requests."
Looking ahead, Hall-Johnston said the pace of change is continuing to increase. And that means organisations will need to remain vigilant to ensure they are not left behind their competition. With Zendesk able to integrate with different platforms through APIs, MCPs and other methods, he expects to see more personal agents being used.
These new products and services are delivered through Zendesk's outcome-based pricing model. Rather than charging for AI usage, Zendesk charges based on the nature and complexity of the resolved outcome, meaning customers only pay when a resolution is successfully delivered.
Customer service should be judged by results and not by the number of tickets closed. By pairing AI agents that learn from every resolved issue with people who step in where empathy and judgement matter, and by charging only when an outcome is achieved, the Outcome AI Platform aligns customer service outcomes with customer expectations. Organisations will need to ensure their systems are integrated, their processes are effective and human expertise remains available when automation reaches its limits.