What Is NiCE CXone? How It Works & Key Features

Ask what is NiCE CXone, and the answer depends on the year the buyer signed.

NiCE CXone is a cloud contact center platform that routes calls and digital messages, answers routine requests with IVR and AI agents, and bundles workforce management, quality management, and analytics in one tenant.

It grew out of NiCE's 2016 inContact acquisition and handles 25B+ interactions a year. Here's how the pieces fit, and where the AI agents on it need testing.

TL;DR: What is NiCE CXone?

• NiCE CXone is a cloud contact center platform that routes voice and digital contacts, runs IVR and AI agents, and bundles workforce management, QA, and analytics.
• NiCE owns CXone. It grew out of the inContact acquisition that closed in November 2016, which is why sign-in and API domains still read "niceincontact."
• The AI layer has carried three names since 2020 (Enlighten, Mpower, and NiCE Cognigy). Every AI agent in a CXone call flow needs the same regression testing as the IVR in front of it.

What is NiCE CXone

NiCE CXone is a cloud platform for running contact center and customer service operations across voice, digital messaging, and AI self-service.

It routes each contact, answers what it can with IVR or AI agents, and hands the rest to human agents. Then it records, scores, and forecasts from every interaction.

NiCE puts CXone volume at 25B+ interactions a year, and its October 2024 count showed 25,000+ organizations in more than 150 countries, including over 85 of the Fortune 100, according to a press release.

The appeal for buyers is one shared record of every contact. The call recording, QA score, schedule adherence, and AI agent transcript all sit in one tenant.

NiCE CXone ownership and the inContact history

The platform traces back to inContact, a Salt Lake City cloud contact center vendor. NiCE announced the acquisition on May 18, 2016, for $14.00 a share in cash.

The deal closed on November 14, 2016. By mid-2017, the combined company had introduced CXone as one cloud platform for routing, workforce optimization, analytics, and AI, hosted on AWS. That launch aimed the enterprise sales push at companies replacing on-premises contact center equipment.

Legacy niceincontact sign-in and help domains

The inContact name never left the infrastructure. Agents sign in at cxone.niceincontact.com, and developers build against the DEVone portal on the same domain.

NiCE shortened the name to NiCE CXone in May 2021, and a June 2025 brand refresh brought the lowercase "i" styling. NICE inContact, NICE CXone, and NiCE CXone on one vendor contract all refer to the same product line.

How NiCE CXone works

Every contact moves through four stages, and each stage maps to a CXone module:

1. A contact arrives: A call hits a CXone number, or a message lands on a digital channel such as WhatsApp, Apple Messages for Business, SMS, or email. A Bring Your Own Channel option covers messaging apps CXone doesn't support natively.
2. Routing picks a destination: The ACD matches the contact to a skill. Each agent holds a proficiency from 1 (highest) to 20 (lowest) per skill, and bullseye routing widens the pool one level at a time when nobody at the top level is free.
3. Automation or a person handles it: A Studio flow runs the IVR menu, an AI agent resolves what it can, and the rest reaches a human in the CXone agent desktop or inside a CRM.
4. The interaction feeds the back office: Recording captures it, quality management scores it, analytics mines it, and workforce management counts it toward the next forecast.

Key features of NiCE CXone

CXone's features fall into four capability areas. Routing, workforce engagement, analytics, and integrations each map to a set of named products.

Omnichannel routing and IVR self-service

CXone Dynamic Routing handles customer-agent matching across voice and digital queues, using each agent's skill proficiency from 1 to 20. The IVR module covers automated phone self-service and call steering.

Both are configured in CXone Studio, a visual flow builder that needs no programming skills. A supervisor can edit a routing flow for every channel from one canvas. Because a business user can change a live flow without a developer, every Studio change needs IVR testing.

Workforce engagement management

CXone's workforce engagement management (WEM) suite covers the human side of the operation:

• Workforce Management: AI-based forecasting and scheduling tied to service levels
• Quality Management: AI evaluation of 100% of interactions, plus coaching workflows
• Recording Management: capture of every interaction for regulatory needs
• Performance Management: agent performance tracking with coaching and gamification

Automated scoring of every call replaces the old habit of reviewing a small random sample by hand. Picking which scores to act on is still your job, and our guide to call center quality assurance software is a good place to start.

Analytics and reporting

Interaction Analytics applies AI to every interaction to find trends worth fixing. Reporting and dashboards add real-time and historical views, with hundreds of pre-built reports across digital and ACD channels.

Reporting tells you how many calls abandoned in the 9 a.m. interval, and conversational analytics tells you what those callers were trying to do before they hung up.

CRM integrations, APIs, and the CXexchange marketplace

CXone embeds inside the CRM your agents already use. CXone Agent for Salesforce lets agents handle voice and digital contacts, plus their schedules, without leaving Salesforce. Equivalent CXone Agent versions exist for ServiceNow and Microsoft Dynamics.

For custom work, the DEVone portal lists hundreds of APIs and SDKs. The CXexchange marketplace carries hundreds of partner solutions.

NiCE CXone AI layer across Enlighten, Mpower, and Cognigy

CXone's AI has carried three names since 2020, and each one added a layer on top of the last. The name on your contract tells you which generation you bought.

Enlighten AI models

Enlighten launched in 2020 as NiCE's set of purpose-built CX models. It uses machine learning, natural language processing, and generative AI to read context, predict intent, and serve real-time recommendations across voice and digital channels.

In March 2024, Enlighten was processing 100 million+ interactions per month.

CXone Mpower copilots, Autopilot, and AI agents

CXone Mpower arrived on June 11, 2024. It combined CXone with three AI tools:

• Copilot: AI assistance for employees across their daily workflows
• Autopilot: virtual agents that detect customer intent and resolve requests directly
• Actions: proactive recommendations that optimize service processes

Autopilot interactions rose 400% in 2024, and CXone handled 6 billion AI-augmented interactions a month that year.

In June 2025, NiCE added CXone Agents (launched as CXone Mpower Agents), AI agents generated from no-code prompts in AI Studio. They run inside orchestrated workflows across front, middle, and back office, handling tasks like processing claims and approving refunds.

Newer NiCE pages drop the Mpower label. The current platform page describes CXone and NiCE Cognigy, so expect older contracts and newer docs to name the same features differently.

NiCE Cognigy AI agents

NiCE announced the Cognigy deal in a press release on July 28, 2025, at roughly $955 million, and closed it on September 8, 2025. Cognigy builds conversational and agentic AI agents for voice and chat.

Cognigy now powers CXone's self-service AI agents. NiCE lists more than 1,000 brands on the product, over 1 billion customer requests a year, and up to 25,000 concurrent conversations on chat or voice.

NiCE reported AI ARR of $362 million in Q2 2026, 15% of its cloud revenue. If you're scoping AI agents on CXone this year, Cognigy is the product NiCE puts forward for self-service.

Who uses NiCE CXone and what it replaces

CXone targets mid-market and enterprise contact centers running several channels, several sites, or regulated workloads. Over 85 of the Fortune 100 run on NiCE, per its October 2024 count.

It usually replaces a stack of separate tools:

• On-premises ACD and PBX hardware that routes calls from a server room
• Standalone workforce management that pulls volume from a separate ACD export
• Separate call recording and QA scorecards reviewed in spreadsheets
• A legacy IVR with touch-tone menus scripted in its own tool
• Point analytics tools bolted onto recordings after the fact

Who should skip it? A small, single-channel support desk pays for routing tiers, WEM, and analytics it won't use. If your real need is an AI agent that answers calls, a dedicated conversational AI platform is a lighter start.

NiCE CXone pros and cons

Pros

• One platform for the whole contact lifecycle: routing, IVR, WEM, analytics, and AI agents share one tenant and one data model.
• Fine-grained skills routing: 1-to-20 proficiencies with bullseye expansion give supervisors exact control over who gets which call.
• No-code flow building: Studio lets business users change routing and IVR flows without developers.
• Deep CRM embedding: agents work inside Salesforce, ServiceNow, or Microsoft Dynamics.
• Published list pricing: per-agent suite prices sit on NiCE's own site.

Cons

• Entry price assumes scale: even the cheapest Omnichannel suite bundles multi-channel routing a small desk won't use.
• Two billing models on the top tier: the Ultimate suite runs per agent or per session, which complicates AI budget forecasts.
• Shifting product names: Enlighten, Mpower, and Cognigy docs, contracts, and help sites overlap across three generations.
• Admin load: skills, proficiencies, and Studio flows need a dedicated owner, since business users can change live flows without a developer.

In April 2026, NiCE was the only vendor named a Customers' Choice in Gartner's Peer Insights Voice of the Customer report for CCaaS, rating 4.7 out of 5 from 82 reviews.

How NiCE CXone compares with Genesys Cloud CX, Five9, and Talkdesk

All four platforms run AI agents on top of omnichannel routing, so the real differences are in the AI layer, the channels each covers, and what the entry price actually buys. Every price below comes from the vendor's own pricing page, checked October 2, 2026.

NiCE CXone pricing summary

CXone sells in five suites from $110 to $249 per agent per month, with a $0.25-per-session option on the Ultimate suite and monthly billing in arrears.

A government package starts at $2 per consumer per year. Pricing checked against NiCE's pricing page on September 30, 2026, so verify with NiCE before you commit.

Testing AI agents on NiCE CXone before and after launch

AI agents on CXone sit inside the same Studio flows, skills, and queues as your human agents. A bad routing edit and a bad prompt reach callers the same way, through a phone line.

Even leading voice platforms miss a real share of calls. In Cekura's workflow benchmark, which ran 8 voice agent configurations through the same 82 scenarios three times each, the top-scoring stack passed all three runs on just 75.61% of scenarios. The rest of the field fell from there to 30.49%.

In one run, an agent read a caller's phone number back correctly, then sent a different number to the tool behind it. This shows how a transcript that looks fine can still hide a broken workflow.

1. Test every IVR menu branch with DTMF input

Callers still press keys. A DTMF test presses every option in every menu and checks where the call lands. It catches a mis-wired Studio branch that dumps billing callers into the sales queue before a real customer finds it.

2. Re-run call flows after every Studio or skill change

Keep a regression suite of scripted calls for each flow and run it after every edit. With bullseye routing, one proficiency change can reshuffle call distribution across a whole skill.

3. Load-test AI agents at forecast peak volume

Use your WFM peak-interval forecast as the concurrency target. The AI platform may scale to thousands of conversations, while the CRM lookups and APIs behind your agent can top out much sooner.

4. Run adversarial tests on AI agents for data leakage and prompt injection

An agent that can approve a refund can be talked into one. Multi-turn attacks raise the odds. The X-Teaming research framework reached up to a 98.1% attack success rate across representative models.

5. Score production calls after each release

Quality Management scores human interactions, and AI agent calls need their own checks. Did the agent complete the refund, hand off with context, and stay inside policy?

Track those outcomes daily for the first week after every prompt or model change. Cekura's production monitoring scores each completed production call against the same metrics you use in testing, and sends a Slack alert when a metric drifts.

How Cekura tests AI agents running on NiCE CXone

Cekura dials your agent over the PSTN like a real caller. There's no native NiCE connector today, so testing runs through the same phone numbers your customers call.

Pre-production

• DTMF and IVR testing: simulated keypad presses walk every menu branch in a CXone flow.
• Multi-turn red teaming: adversarial conversations spanning 6 to 10 turns probe for prompt injection and data leaks.

Infrastructure

• Load testing: concurrent test calls expose bottlenecks and latency spikes at your peak volume.
• SIP testing: SIP-based agents get tested over IP with custom headers.

Observability

• Scheduled runs and alerts: cron-based calls check your live agent on a schedule, and Slack alerts fire when a metric drifts.

Agents built outside CXone plug in directly. Native integrations cover Retell, Vapi, ElevenLabs, LiveKit, Pipecat, Bland, and more.

Cekura supports SOC 2, HIPAA, and GDPR compliance, covering PII redaction and role-based access control. Pay-as-you-go pricing starts at $0.25 per voice testing minute, with the first seat free and $30 a month for each additional seat.

So what is NiCE CXone in 2026? A contact center platform whose AI agents now answer calls that used to reach people.

When you skip that testing, the failures stay invisible. You get refunds that never process, handoffs that drop context, answers that drift outside policy after a prompt change. Book a demo to watch Cekura run DTMF, load, and adversarial calls against your CXone numbers.

Frequently asked questions

What is NiCE CXone used for?

NiCE CXone is used to run a contact center in the cloud. It routes calls and digital messages, answers routine requests with IVR and AI agents, and records, scores, and forecasts every interaction.

Is NiCE CXone cloud-based?

Yes. NiCE CXone is a cloud-native contact center platform that has run on AWS since its 2017 launch.

Who owns NiCE CXone?

NiCE owns NiCE CXone. The platform came from NiCE's acquisition of inContact, which closed in November 2016, and it launched under the CXone name in July 2017.

Is CXone the same as inContact?

Mostly, yes. CXone is the cloud platform NiCE built from inContact's contact center after the 2016 acquisition. It was sold as NICE inContact CXone until May 2021, when NiCE shortened the name to NiCE CXone.

The inContact name survives in legacy domains such as the cxone.niceincontact.com sign-in page.

What is CXone Mpower?

CXone Mpower is the name NiCE gave in June 2024 to CXone combined with its Copilot, Autopilot, and Actions AI tools. NiCE's newer pages describe the same platform as CXone with NiCE Cognigy AI agents.

How much does NiCE CXone cost?

NiCE CXone costs $110 to $249 per agent per month across five suites. The top Ultimate suite also offers a $0.25-per-session option, billed monthly in arrears.

Does NiCE CXone include workforce management?

Yes, NiCE CXone includes workforce management with AI-based forecasting and scheduling. It sits in the WEM suite alongside quality management, recording, and performance management.

Can you test AI agents built on NiCE CXone?

Yes, you can test AI agents on NiCE CXone by calling them over the phone or SIP with simulated callers. Cekura runs those calls at volume with DTMF, load, and adversarial scenarios, then scores the outcomes.