NiCE Adds Testing and Tracing to Cognigy AI
The update to NiCE's Cognigy platform adds agent tracing, multivariate testing, and reusable skills, with trial access starting November 12, 2026.
Updated on Nov. 14, 2026 in Artificial Intelligence
NiCE has added testing, tracing, and reusable-skill tools to its Cognigy AI platform, including a redesigned Agentic Experience Studio and AI Assist for building and improving agents. Trial access to the new features begins November 12, 2026.
Why it matters
The update lets teams record the steps, data, tools, and decisions behind agent actions and compare configurations before deployment. NiCE reported its AI and self-service annual recurring revenue reached $362 million as of June 30, with 52% growth from a year earlier.
The platform adds native speech-to-speech support, with default speech-to-text, text-to-speech, and large language models included in the conversation price, while premium models from Deepgram, ElevenLabs, and Anthropic carry usage-based billing. Performance benchmarks for the new features have not been disclosed.
The players
NiCE
NiCE is a publicly traded software company that provides customer experience and contact center technology, and owns the Cognigy conversational AI platform.
The details
Agentic Experience Studio has been redesigned, and AI Assist helps teams set up, build, test, and improve agents. Tracing records the steps, data, tools, and decisions behind agent actions, and a flow can serve as an agent tool or complete a defined step before an agent takes over. New agents can work alongside existing Flows and Flow Agents.
Timeline
June 30, 2026: NiCE's AI and self-service ARR figure is measured as of this date.
November 12, 2026: Trial access to the new features begins.
November 2026: A gradual rollout to customer environments is planned.
The Tech Race
The Cognigy update extends the platform's existing Flows and Flow Agents with tracing, testing, and reusable skills, placing NiCE in the contest among enterprise AI agent vendors to make agent behavior auditable. Competitors in conversational AI are likewise adding observability and evaluation tooling as agents move into production contact-center use.
Teams building customer-service agents can now trace each action an agent takes and compare configurations through multivariate testing before deploying changes. Speech-to-speech support and default speech and language models are included in the conversation price, while premium models carry separate usage-based charges.
The takeaway
Teams evaluating AI agents can use tracing and multivariate testing to see why an agent acted as it did and to compare configurations before rollout. Watch the billing terms for premium voice and language models, since usage-based charges apply beyond the default conversation price.
What happens next
Trial access to the new features begins November 12, 2026, and a gradual rollout to customer environments is planned for November 2026.
Further reading
For more on how companies are building and governing these systems, see Artificial Intelligence coverage.
Source note: This article includes information reported by TokenPost.
