PolyAI has launched its new Agent Development Kit (ADK), aimed at revolutionizing how teams develop, deploy, and enhance agentic AI for customer experience. This innovative toolkit introduces an AI-native development model, enabling developers to utilize coding assistants as integral components in building customer service agents. The approach marks a departure from traditional methods that often rely on static configurations or manual implementations.
According to Shawn Wen, co-founder and CTO at PolyAI, the ADK is designed to align with the workflows of developers. “Most AI platforms for CX force developers to work inside a UI, cut off from the way real software gets built,” Wen noted. “With the ADK, we’re changing that. Developers can now build AI agents with the same tools, workflows, and flexibility they use to build any other critical system.”
The ADK offers several key functionalities that allow developer teams to work more efficiently. They can build in their own environments, utilizing their preferred programming languages and integrated development environments (IDEs) with full visibility into the codebase. The kit integrates AI coding assistants, such as Cursor and Claude Code, to accelerate development processes. Additionally, it enables the creation of agents from various inputs, including diagrams, spreadsheets, and APIs, in a matter of minutes. This capability allows teams to manage AI agents with enterprise-grade software tools, complete with version control, code reviews, and collaborative workflows.
This toolkit has already demonstrated its potential for rapid development among PolyAI’s enterprise customers. For instance, one of the world’s largest utility companies was able to create complex workflows in hours, a process that previously took several weeks. Similarly, a major UK bank managed to process hundreds of frequently asked questions in mere minutes, significantly reducing the manual effort required.
The ADK is also designed with a focus on continuous improvement. By connecting to monitoring tools and automated quality assurance pipelines, AI agents can identify performance gaps and engage in self-improvement without the need for ongoing manual intervention. This represents a significant shift from static automation to dynamic systems that learn and adapt over time.
Wen highlighted an internal benefit, stating that more than 60% of engineering work at PolyAI is now conducted autonomously through ADK-powered workflows, with developers primarily reviewing and guiding the outputs. “Now, we’re making that same capability available to our customers so more teams can build high-quality AI systems faster,” he added.
The introduction of the ADK signifies a broader trend in the customer experience sector, where traditional AI vendors often confine developers to closed systems with limited code access and integration capabilities. In contrast, PolyAI’s approach empowers engineering organizations to utilize the same tools and methodologies they employ for other critical software systems. Combined with the no-code and low-code functionalities in the company’s core Agent Studio platform, PolyAI positions itself as a unique player in the enterprise AI landscape, catering to both technical and non-technical users.
As organizations increasingly seek sophisticated AI solutions to enhance customer interactions, PolyAI’s ADK could set a new standard for building enterprise-grade AI. The ongoing evolution of AI technology continues to reshape how businesses engage with their customers, emphasizing the need for scalable, adaptable solutions that meet diverse operational demands.
