The Gist
AI bets big at Interactions. NiCE launches Mpower Agents at its Las Vegas conference, pushing for enterprise-wide automation with AWS and Snowflake partnerships. From chat to action. Mpower Agents aim to go beyond conversation, driving outcomes across self-service, mid-office, and fulfillment workflows. Reality check from analysts. While NiCE accelerates Agentic AI, many brands remain hesitant to deploy GenAI in customer-facing roles due to legal and trust concerns.
The contact center and customer service software space seems to boil down to which “A” you want to lead with.
Agentic AI. Or, just AI. Or, throw in a “G.” Generative AI.
NiCE, a contact center as a service provider, recently rebranded itself as an “AI company.” It is betting big on AI and Agentic AI, really. Betting’s appropriate. The provider holds its Interactions conference this week at the Aria Resort & Casino.
NiCE today announced the launch of CXone Mpower Agents, which it says enables fully automated, AI agents to be “created and deployed in seconds and immediately work across the customer service ecosystem — from self-service to mid-office approvals to back-end fulfillment.”
Mpower Agents leverage CXone Mpower’s proprietary CX AI models, which are trained on deep, use case-specific data and refined through workflows modeled after top-performing employees. These agents operate across systems, initiate actions and collaborate with both humans and other AI agents to fully automate service workflows, according to company officials.
“There’s a big difference between AI that talks and AI that gets things done,” Barry Cooper, president, CX Division, NiCE, said in a release. “While others are building agents that mimic conversations, we’re building agents that fulfill customer needs — end to end. Whether it’s a mid-office approver or a back-office loan processor, Mpower Agents work across the entire CXone Mpower platform to deliver real outcomes, not just responses. That’s what separates intelligent automation from intelligent distraction.”
Table of Contents
NiCE Puts Automation at the Center With Mpower Agents
NiCE rolls the dice on automation because unlike traditional automation approaches that require heavy coding and complex builds, Mpower Agents from NICE are designed with automation at the center from day one, according to company officials. Rather than manually developing each solution, the CXone Mpower platform uses built-in intelligence to spot high-value automation opportunities across the entire customer operation — front office, middle office and back office.
Mpower AI Studio allows users to create these agents using outcome-based prompts—no coding necessary. Once created, these agents are deeply integrated with the full CXone Mpower ecosystem, including:
APIs for system connectivityKnowledge bases for information accessExperience Memory for contextual learningChannels for communication across platformsEnlighten AI Models for behavior-driven intelligence
These agents can be deployed in two main ways:
Mpower Copilot to support employees behind the scenesMpower Autopilot to engage directly with customers
“AI agents are becoming essential for modern customer service, but most still fall short, limited to scripted responses or narrow front-office use cases,” Maribel Lopez, principal analyst at Lopez Research, said in a NiCE release. “What businesses need are solutions that provide the ability to use automated insights to identify opportunities and instantly create agents that operate across front, mid and back office. NiCE’s Mpower Agents aim to solve previous issues by focusing on intelligent automation.”
Related Article: NiCE & Snowflake Partner to Unify CX Data Across Enterprise
NiCE Product Releases and Capabilities
Summary of the major NiCE product announcements and enhancements revealed at the 2025 Interactions event.
Product/CapabilityDescriptionStrategic ImpactCXone Mpower AgentsAI-powered agents created in seconds, designed to automate full service workflows across front, mid, and back office.Accelerates intelligent automation across the entire customer journey.Mpower AI StudioNo-code environment for generating outcome-based AI agents using simple prompts.Democratizes AI agent creation for business users, reducing reliance on developers.Mpower CopilotReal-time agent assistance tool for employees, supporting decisions and tasks behind the scenes.Improves productivity and decision-making across service roles.Mpower AutopilotCustomer-facing AI that engages directly across channels to resolve issues autonomously.Expands automation to self-service and reduces agent load.Amazon AI IntegrationPartnership with AWS integrating Amazon Bedrock, Q, SageMaker and Nova LLMs into CXone Mpower.Boosts scalability, context-awareness, and enterprise-grade performance of Mpower agents.Snowflake CollaborationPartnership to connect CXone Mpower with Snowflake’s Data Cloud, enabling secure, real-time access to enterprise-wide customer data.Eliminates data silos, enhances personalization, and allows AI agents to act on unified insights across departments.Experience Memory + Enlighten ModelsBehavior-driven AI models that learn from top-performing employee interactions and CX history.Improves agent training, accuracy and personalization of AI responses.
Push for Agentic AI Is Real, While Some Brands Remain Cautious
The Agentic AI push is here. No doubt.
Is that the right play for leading providers like NiCE?
Max Ball, principal analyst, Forrester, highlights the tension between vendor hype and real-world hesitation when it comes to generative AI in contact centers. While solution providers race ahead with flashy claims about “agentic” AI, many brands are still hesitant to roll out generative AI for customer-facing roles — largely due to legal concerns and trust issues.
“Brands are being very careful and frustratingly slow to adopt GenAI for direct customer facing apps (IVAs and chatbots),” Ball told CMSWire. “When I talk to brands one of the answers I get is fear/discomfort with the uncontrollable nature of GenAI apps. It’s one thing to say, ‘This bot will always say this when answering this question.’ It’s quite another to say, ‘Well, I’m not totally sure what it will say and while we have done everything we can do to reduce hallucinations to the point where we believe it is more reliable than a human agent, I can’t guarantee there will be no mistakes.’”
As one contact center manager put it to Ball: “We can’t get this past the lawyers.”
“I do think there is a lot of that going on right now,” he added.
Related Article: Navigating AI Hallucinations: A Personal Lesson in Digital Accuracy
Enterprise Adoption of Agentic AI Is Gaining Ground
We’re hearing more and more about “Agentic AI” and orchestration layers. Are most contact centers ready for that complexity, or is it still early days?
It’s definitely happening.
Deloitte estimates that by 2025, one in four companies using generative AI will begin testing Agentic AI through pilots or proof-of-concept initiatives—a number expected to rise to half of such companies by 2027. While widespread deployment remains on the horizon, certain industries and use cases may start integrating Agentic AI into existing workflows as early as the second half of 2025.
Gartner projects that by 2029, Agentic AI will independently handle 80% of routine customer service inquiries, cutting operational costs by 30% through reduced reliance on human agents.
How’s it looking in practice?
“I think I’m in the minority when I say that most contact centers aren’t ready for Agentic AI,” Blair Pleasant, president and principal analyst, COMMfusion LLC & BCStrategies, told CMSWire. “I still think about all of the contact centers that haven’t deployed screen pop, which has been available since the ‘90s, so it’s hard to imagine that they’re ready for Agentic AI. Of course, there are early adopters that can take advantage of Agentic AI, and we’re hearing about some early success stories, but for the most part, I think it’s still very early days.”
Learning OpportunitiesView all
Much of what vendors are calling “agentic” is just “IVAs on steroids,” she added, and not truly agentic, which includes reasoning and action, and the ability to perform tasks on behalf of someone, or something.
“We’re starting to see vendors introduce true agentic capabilities,” Pleasant added, “but it will take time for the majority of enterprises and contact centers to truly embrace it.”
Vendors in general—and not singling out NiCE—have quickly shifted their messaging from generative AI to so-called “agentic” solutions, which became a buzzword at this year’s Enterprise Connect, according to Forrester’s Ball. But the reality is, while contact centers are experimenting with GenAI, they’re avoiding its use in customer-facing roles—where it could have the biggest impact.
Some vendors overpromise on agentic capabilities that remain limited and overly constrained, falling short of what most would actually consider agentic. Ball finds it frustrating that marketing on Agentic AI is racing ahead of what may be the fastest tech rollout in history (generative AI), while customers remain hesitant to take even the first practical steps.
Related Article: Agentic AI: A New, Global Imperative for Customer Experience Leaders
Where the Contact Center Industry Is Going With AI
Key trends, analyst insights and projections shaping the future of AI in contact centers.
TrendInsight/StatisticSource/ContextSlow GenAI AdoptionBrands remain cautious about deploying GenAI in customer-facing roles due to legal risks and hallucination concerns.Forrester (Max Ball); CMSWire interviewAgentic AI Momentum25% of GenAI-using companies expected to pilot Agentic AI in 2025; projected to rise to 50% by 2027.Deloitte, 2025 TMT PredictionsAutonomous Service by 2029Agentic AI will handle 80% of common service issues, reducing operational costs by 30%.Gartner, 2025 PredictionDefinition GapMany “agentic” solutions are still IVAs with enhancements—not truly autonomous agents with reasoning and execution.COMMfusion (Blair Pleasant); CMSWire interviewCustomer Familiarity RisingAI assistant usage is growing, but companies must still guide customers through the transition.McKinsey, Contact Center Crossroads reportTransactional Volume Still High50–60% of calls are still basic transactional inquiries, showing opportunity for automation.McKinsey, 2025 Operations Insights
NiCE Doubles Down on Agentic AI With AWS Partnership
But NiCE, for one, isn’t slowing down the Agentic AI push.
NiCE also announced today it expanded collaboration with AWS is positioned as a major step toward accelerating Agentic AI, combining CXone Mpower with Amazon’s generative AI infrastructure—including Bedrock, SageMaker and Amazon Q. The partnership promises enterprise-wide automation, real-time orchestration and instant AI agent creation without code. On paper, this signals a fast track to intelligent customer service at scale.
The integration with AWS enables CXone Mpower to tap into Amazon’s Nova family of large language models (LLMs), offering scalability, speed and enterprise-grade security for AI deployments. Through this collaboration, businesses can deploy intelligent automation across front, middle and back office operations — streamlining tasks like claim processing, refund approvals and customer query handling through orchestrated AI workflows.
Additionally, NiCE is enhancing its AI agent capabilities with tools like Amazon Q index and Amazon SageMaker, allowing agents to access up-to-date business content and learn from historical performance data. The result is a more context-aware, responsive system that supports both employee augmentation via Mpower Copilot and direct customer interactions via Mpower Autopilot, aiming to deliver measurable gains in productivity and customer satisfaction.
Agentic AI or Generative CX? The Road Ahead
So now it’s about where providers like NiCE will go next: will Agentic AI be the heavy push through the rest of 2025 and into 2026? Will vendors in the contact center space get back to more “generative AI in CX” messaging?
Remember, we’re talking about a space where some claim the “majority of transactions are still simple.”
McKinsey & Company’s review of millions of customer interactions across over 30 organizations reveals that although the complexity of calls has grown, transactional inquiries still make up 50% to 60% of all interactions—despite major efforts to reduce them. At a leading European bank, for example, simple requests like checking recent transactions or paying bills still represent half of all call volume. Similarly, a North American telecom provider reports that about 40% of its calls are transactional, such as refund requests, plan inquiries or new device questions.
At the same time, it can’t be discounted that customer familiarity with generative AI is increasing, thanks to growing use of AI assistants for tasks ranging from drafting emails to solving more complex problems. This rising comfort with AI could encourage broader acceptance of AI-powered customer service tools—or compel businesses to deploy them to meet customer expectations, McKinsey & Company reported.
However, full adoption may not happen without a push. Organizations will likely need to guide customers through the transition, identify barriers to use and build tailored strategies to accelerate uptake, researchers noted.
From CCaaS to CX-First AI: Inside NiCE’s Rebrand
NiCE is all in on AI, though.
Under new CEO Scott Russell, the company is shifting from a traditional CCaaS provider to an AI-first organization—one that still places strong value on the “human touch,” according to Omer Minkara, Analyst at Aberdeen Research. “While NiCE empowers business leaders with automation, generative AI, and Agentic AI to execute customer experience processes, it also seems to recognize and leverage the human side of AI,” he wrote.
The rebrand acknowledges both the promise and the limitations of generative AI, particularly its susceptibility to hallucinations. To mitigate these risks, NiCE is advocating for a model where humans augment AI—not just execute tasks, but supervise and validate them to ensure quality, Minkara added. This positions AI-proficient employees as critical to realizing AI’s value while safeguarding CX integrity. As the company evolves, NiCE is expected to expand into adjacent markets like sales, marketing and ecommerce—extending its CX-first AI strategy beyond the contact center, he noted.
Everyone is trying to differentiate with AI right now, according to Pleasant, who added that NiCE is focusing on several areas:
AI and automation with a human touch, using AI to empower agents and employees. They’re focused on amplifying human experiences, where technology doesn’t replace people, it amplifies them.Being proactive and resolving issues before the customer even knows there’s a problem.
“The company is trying to move beyond the contact center,” Pleasant told CMSWire. “The new CEO comes from the larger enterprise space, and wants the company to make a mark outside of the contact center and focus more on the C-suite and AI. The new branding is part of an effort to expand beyond CCaaS to intelligent, AI powered, world-class end-to-end service automation.”
Have a tip to share with our editorial team? Drop us a line:
