NiCE Cognigy adds reusable agent skills and decision tracing

By Sophie Lindqvist

On Nov. 12, NiCE launched the next generation of NiCE Cognigy, an enterprise AI-agent platform, with AI-assisted development, reusable skills, decision tracing and native integrations from Deepgram, ElevenLabs, OpenAI and Anthropic. Existing production deployments continue unchanged.

The platform combines autonomous agents with structured flows and governance controls. Teams can decide where an agent acts independently and where a defined process governs the work. NiCE says the release moves enterprises beyond automating individual tasks toward orchestrating AI agents across entire customer journeys.

What changes for existing Cognigy customers?

Existing customers receive the next-generation capabilities at no added cost as part of NiCE's continued product work. Live deployments keep running without modification, and customer teams can start using the new features when ready. The default stack needs no configuration and carries no separate model contract.

The release follows NiCE's $955 million acquisition of Cognigy, which closed Sept. 8, 2025. The deal embedded Cognigy's conversational AI in the CXone Mpower platform and extended NiCE's reach to clients including Mercedes-Benz, Lufthansa Group, Nestlé and Toyota.

A faster studio and reusable building blocks

The redesigned Agentic Experience Studio gives teams an AI-assisted workspace for building agents. NiCE says it shortens the path from an idea to a working agent. New projects start with default speech-to-text (STT), text-to-speech (TTS) and large language models, plus ready-made resources.

AI Assist guides setup, applies best practices, and supports teams as they build, test and improve agents over the agent lifecycle.

NiCE Chief AI Officer Philipp Heltewig said: “Great AI Agents come from teams that can spend their time on what the agent does for customers. We designed the next generation of NiCE Cognigy around that, so enterprises can scale AI Agents with the visibility and control production demands.”

Each agent uses skills, tools and personas. Teams create these components once, reuse them across agents and improve them over time. For complex work, an agent can coordinate specialized sub-agents.

Agents can also connect with external agents through open standards, including a customer's personal AI agent. NiCE designed that connection for companies whose customers now arrive with an AI agent of their own.

Tracing records every step an agent takes, including the data and tools it uses and the decisions it makes. Heltewig said: “AI Assist guides setup, testing and improvement. Skills built once work across every agent. Tracing shows every decision.”

The release also adds native speech-to-speech models that process audio end to end. They work on the audio directly rather than converting speech to text and back. NiCE says this produces faster, more lifelike conversations.

Teams can add premium STT, TTS or LLM models from Deepgram, ElevenLabs, OpenAI or Anthropic in a few clicks, with usage-based billing. Teams with their own model contracts can continue using them. Default models remain included in the conversation price.

NiCE's deployment and financial context

NiCE launched an AI Ops Center in October 2025 and a Cognigy Simulator in January 2026. At NiCE World 2026, the company declared agentic AI native to its architecture. Gartner named NiCE a Leader in its 2026 CCaaS Magic Quadrant and placed it furthest for Completeness of Vision.

Q3 2026 revenue reached $790.7 million. AI ARR rose to $392 million from $345 million in Q1. NiCE reiterated full-year guidance of $3.17 billion to $3.19 billion. Partnerships with Snowflake, Salesforce, AWS and ServiceNow deepened, and NiCE agreed in July to resell RingCentral's RingEX.

In November, NiCE disclosed plans to cut about 500 jobs, roughly 5% of its workforce. The company expects $15 million to $20 million in restructuring charges in Q4 as it reorganizes toward enterprise-scale CX AI. Chairman David Kostman will transition from his role on Dec. 3, 2026.

The market evidence for scaled agents

Enterprise AI agents have moved from pilots to production, with documented improvements in resolution rates, handle time and costs. The 2025 CMSWire State of the CMO Report found a 28% improvement in issue resolution time and a 19% increase in first-contact resolution among organizations using autonomous AI systems.

Additional CMSWire benchmarks show:

• 40% of inquiries resolved autonomously across chat, email, voice and WhatsApp.
• 14% more issues resolved per hour with generative AI-enabled agents.
• 20% to 30% lower contact center costs through omnichannel AI deployments.

Production-grade deployments increasingly pair AI autonomy with business-logic constraints. CMSWire reporting cautions that curated data and manual workarounds can mask coordination failures in pilots, which emerge only at scale.

Gartner predicts that agentic AI will independently handle 80% of routine service inquiries by 2029, cutting operational costs by 30%.

NiCE provides enterprise software for customer service, fraud prevention and compliance. Founded in 1986, the company trades on Nasdaq. Its CXone platform combines self-service, omnichannel routing and workforce-management tools, using AI to automate interactions and support agents.