NiCE Rebuilds Cognigy as an Agentic-First Platform

Author: Beth Schultz, VP of Research & Principal Analyst

What's the news?

On Nov. 12, 2026, NiCE launched an updated version of its NiCE Cognigy conversational and agentic AI platform, reworking the product around AI agents rather than the natural language understanding (NLU) intents and deterministic flows on which Cognigy built its reputation. With this release, NiCE aims to shorten the path from idea to production-ready AI agent while giving builders visibility into how those agents reason and act. The new capabilities include:

• Agentic Experience Studio – offers a redesigned, AI-assisted build environment that comes with default speech-to-text (STT), text-to-speech (TTS), and large language models (LLMs), plus ready-made resources, so teams can stand up an agent without first configuring models
• Next-generation AI agents – assembled via reusable skills (packaged instructions and tools organized around a task or theme), tools, and personas; these agents can coordinate specialized sub-agents and interact, via open standards, with external agents, including consumers’ personal AI agents
• AI Assist – embeds an AI companion across the agent lifecycle to guide setup, apply best practices, and help build, test, and refine agents
• Tracing – exposes each step an agent takes, from the data and tools it draws on to the decisions it makes
• Voice-to-voice support – adds native speech-to-speech models that process audio end to end rather than converting it to text and back

NiCE positions the platform as hybrid by design, letting customers pair autonomous agents with structured flows, safeguards, and governance controls. Existing deployments can keep running unchanged, with current customers receiving the next generation as part of their ongoing NiCE Cognigy service, adopting the new capabilities on their own timelines, NiCE said during an analyst briefing.
NiCE did not disclose pricing. It said the conversation price covered default AI models, with no separate model contract required, while premium STT, TTS, and LLM models from Anthropic, Deepgram, and ElevenLabs were available with usage-based billing; customers can also retain their own model contracts.