Zendesk Revolutionizes Customer Service with Autonomous AI

Zendesk and the Era of AI Agents

In 2025, Zendesk introduced its Resolution Platform, marking a significant advancement in the use of AI for customer service. This year, the company is taking it a step further with the announcement of the Autonomous Service Workforce. This initiative aims to radically transform the way businesses interact with their customers by using AI agents to automate and assist in numerous tasks. The idea is to enable companies to perform actions that were previously economically unfeasible.

The rapid evolution of AI has prompted Zendesk to adapt its vision. A year and a half ago, the company was already exploring the use of AI agents to enhance customer conversations. Today, with the Autonomous Service Workforce, every aspect of work is influenced by AI, whether it involves assistance or automation. The Custom Zendesk Agents open up new use cases, allowing for cognitive reflections and capabilities that were once reserved for humans.

Humans remain a valuable resource, and thanks to AI, many tasks that were previously unachievable are now possible. The Autonomous Service Workforce not only automates existing tasks but also allows for a rethinking of what is possible and the construction of new solutions.

Acquisition Strategy for Innovation

Zendesk has adopted an aggressive acquisition strategy to bolster its AI capabilities. Since its acquisition by Hellman & Friedman and Permira for $10.2 billion in 2022, the company has integrated several key firms. Among them are Klaus for quality assurance, Tymeshift for workforce management, and HyperArc for data analytics. These acquisitions aim to bring innovation, talent, and valuable time savings.

The acquisition of Forethought in 2026 is particularly noteworthy, as it has been directly integrated into Zendesk's AI agents, offering continuous improvement through the Resolution Learning Loop. This intelligence system analyzes each interaction and provides recommendations to optimize processes. For example, it can automatically adjust product return procedures based on shipping rules specific to each region.

Launches and Innovations

Zendesk has also launched several new platforms, such as the Custom Agents Platform, ITAM (IT Asset Management), and the AI Actions Platform for integration. These innovations represent what Zendesk calls "zero to one" innovation, where the company builds solutions from scratch after assessing the market.

In parallel, Zendesk has introduced the MCP Client and the MCP Server to align with market standards. These tools are designed to be game-changers for clients, although the impact varies between the Client and the Server.

Zendesk's AI Acquisitions

Since 2022, Zendesk has made several acquisitions in the AI space:

• Klaus (2023): Quality Assurance, now Zendesk QA
• Tymeshift (2023): Workforce Management
• Local Measure (2024): Voice and contact center
• Ultimate (2024): Historical AI Agents
• Unleash (2024): Enterprise search, foundation for AI Agents for Employee Service
• HyperArc (2025): Analytics, foundation for the Context Graph and Analyst Copilot
• Forethought (2026): AI-native AI Agents, directly integrated into Zendesk AI Agents and now deployable across multiple platforms.

Continuous Improvement with the Resolution Learning Loop

To truly achieve continuous improvement, one must leverage what is available: an intelligence system that examines every interaction, every piece of data, and can formulate recommendations to enhance service. Take the example of a retailer whose AI agents fail to automatically process product return requests in Canada when the product contains a liquid, as shipping rules prohibit it. With the Resolution Learning Loop, when such a case arises, the system can automatically generate the new procedure for the AI and present it to the admin: “Would you like to approve this, or make slight modifications? We believe we could provide better service by doing it this way.”

This is what the knowledge manager or the service operations manager has been doing until now: gathering feedback from one agent, cross-referencing it with another, identifying a pattern. But each customer service agent only sees one ticket at a time. You may have one experience, and someone in Istanbul may have exactly the same, without ever being connected. To do this at the scale of a large company would require a managerial organization that does not exist. Therefore, we want to build systems that do this in place of humans, so that companies can continuously improve and capture every signal.

The Concept of "Service Debt"

The concept of "service debt" is essential for understanding how companies can continuously improve. It refers to the debt accumulated by inefficient or outdated processes, which Zendesk's AI aims to resolve by automating and optimizing customer service operations.

The Proliferation of Specialized AI Agents

Tom Eggemeier discussed the proliferation of specialized AI agents, such as customer success agents, community agents, and support agents. This diversification of roles means that human teams in customer service must adapt to these new dynamics. The increase in the number of AI agents does not necessarily mean fewer humans, but rather a reconfiguration of roles, where AI and humans work together to provide enhanced customer service.