What is an AI-Native Contact Center? Architecture vs Add-Ons
An AI-native contact center is one where AI is built into the core architecture from the ground up rather than layered on top of existing infrastructure. That distinction, largely invisible to buyers evaluating platforms, was the central argument Chris Morrissey, General Manager of Zoom Customer Experience, made at CCW Vegas 2026.
We’re not AI-first, we’re a CX-first business
“We’re not AI-first,” Morrissey told Rob Scott. “We’re a CX-first business. If the technology changes, it changes.”
It is a deliberately counterintuitive position for a company actively deploying AI across its entire contact center stack. But it goes to the heart of a debate that is quietly dividing CX leaders right now and Morrissey’s argument, made in full at CCW 2026, is worth unpacking in detail.
What AI-native architecture actually means — and what bolted-on AI looks like in practice
The term “AI-native” is now applied so broadly that it has lost most of its meaning. For Morrissey, the more useful question is not whether a contact center platform claims AI-native status, but whether its AI components can actually communicate with each other.
“Historically, data silos have been the big challenge,” he explains. “We could now get into a situation where we have AI silos instead. One AI acting as a copilot for agents through Expert Assist, a different AI for your voice virtual agent, a different AI for your chat agents — you’re going to have different experiences across all of those.”
The moment those silos surface for customers is specific, and Morrissey describes it with precision: a virtual agent handles the first part of a call, the customer escalates to a human agent, and the human agent has no visibility into what the virtual agent already tried. The agent asks the customer to repeat everything. The customer is frustrated — not despite the AI investment, but because of how it was deployed.
By not connecting the journey, you’ve used AI — but you’ve also created friction for the customer. You’ve made the experience worse.
Zoom CX’s answer is a platform where quality management data, interaction data, workforce management data, and the full conversation history of both human and virtual agent interactions feed a single AI layer. That shared context is what enables the platform to self-improve over time — making both virtual agents and human agents incrementally better with each interaction. It is the difference, Morrissey argues, between genuine platform integration and what he calls “a unified front end” masking a fragmented back end.
“Make sure you know what unified really means,” he says. “A lot of vendors, if they’re honest, it’s actually a unified front end.”
AI-Powered Contact Center Tools: Outcomes Zoom CX Customers Are Reporting
AI-powered contact center tools are delivering measurable improvements across resolution rates, CSAT scores, and agent performance — but only when deployed against clearly defined business outcomes. Zoom CX customers are reporting gains across all three, with the shift from deflection metrics to resolution metrics marking the clearest indicator of genuine progress.
From deflection rates to resolution rates: why the metrics are changing
The boardroom conversation around AI ROI has, for years, defaulted to efficiency: handle time reduced, calls deflected, headcount avoided. Morrissey is direct about why that framing is no longer sufficient.
“The concept has shifted — it’s not about deflecting calls, it’s about resolving them,” he says. “Resolving calls with a virtual agent. Helping human agents with guidance and Expert Assist so they can do their jobs better. And then giving the business better insights through tools like CX Insights — tell me something I don’t already know.”
To hold itself accountable to that standard, Zoom CX has built resolution dashboards inside its Virtual Agent product that use AI to evaluate AI performance — always with a human in the loop. The dashboards track resolution rates, whether those rates are improving, and the human cost of each call. That data is then used to demonstrate real value to customers: not just cost savings, but what those savings are being reinvested into.
The IKEA deployment is the example Morrissey returns to. The retailer redirected approximately 8,500 roles through AI adoption but created new ones rather than eliminating positions.
IVR and Virtual Agents: How Zoom CX Handles Automated Customer Interactions
Zoom CX’s Virtual Agent is designed to resolve interactions, not just deflect them. It runs on the same AI and data layet as the human agent desktop, so the full conversation history, including what the customer asked, what the virtual agent attempted, and where it fell short, carries forward automatically at handoff. When a customer escalates from virtual to human, the agent sees exactly what happened in the prior interaction, eliminating the need to repeat diagnostic steps and the frustration that typically accompanies that transition.
The platform moves beyond intent-based IVR toward fully agentic AI capable of handling complex, multi-turn interactions. Morrissey notes that the shift from intent-based to agentic AI happened faster than most of the industry anticipated. Resolution performance is monitored through built-in dashboards, with human oversight ensuring quality at scale and AI is used to continuously evaluate its own outputs.
Real-Time Agent Coaching and Contact Center Quality Management: How Zoom CX Uses AI
Real-time agent coaching and contact center quality management (QM) are two of the areas where Zoom CX’s unified data architecture creates the clearest competitive advantage. Because quality management data, call data and workforce management data all feed the same AI layer, the platform can use what it learns from every interaction to improve what happens in the next one across both AI-handled and human-handled conversations.
Expert Assist is the agent-facing expression of that. During live interactions, it surfaces guidance in the background, such as relevant knowledge articles, suggested responses and next-best actions. Surfacing these in the background without interrupting the agent’s conversation with the customer allows the agent to stay focused. The AI handles the retrieval and the routing. Morrissey frames the goal not as replacing the agent’s judgment but removing the operational weight that gets in the way of it.
CX Insights takes the same logic upstream to supervisors and business leaders. “It used to be reports, then dashboards,” Morrissey told CX Today. “And now insights are better than dashboards – tell me something I don’t already know.” Where traditional QM tools require manual report-pulling across systems, CX Insights surfaces patterns, trends and answers from across the entire operation. This business intelligence reflects the whole contact center, rather than snapshots of individual channels.
Underpinning both is the resolution dashboard layer built into Zoom CX’s Virtual Agent, using AI to evaluate AI performance. But humans are always in the loop: tracking resolution rates, improvement trends and the real cost of each interaction. Quality management, in Zoom CX’s architecture, applies not just to human agents but to the entire AI-powered stack.
Omnichannel Contact Center Tools: Voice, Video, Chat and Messaging in One Platform
An omnichannel contact center platform gives customers a consistent experience regardless of the channel they use, voice, video, chat, or messaging, while giving agents a single surface from which to manage all of them. Most vendors claim this capability. Fewer deliver it at the architectural level.
Voice, video, chat, and messaging in one surface — and why the agent experience is the real differentiator
The challenge with most unified platform claims, Morrissey argues, is that they describe a front-end experience rather than a genuine integration. “When a lot of vendors say it’s a unified platform, if they’re honest, it’s actually a unified front end. There’s a lot going on in the background to make it look that way.”
A unified front end may present a single interface to agents while routing data across multiple disconnected back-end systems. A genuinely unified platform shares context across every channel in real time, without requiring manual retrieval from conversation history, customer data, or prior interaction outcomes
For Zoom CX, omnichannel unification is built around the agent experience. The platform supports voice, video, chat, and messaging from a single desktop, with AI surfacing customer history, prior interaction and conversation intent in real time as the channel changes. It is the same AI layer that powers virtual agent interactions, which means the context that accumulates in automated interactions carries forward into human ones.
Contact Center AI Solutions: How Zoom Keeps Agents in Control
Contact center AI solutions that keep agents in control, rather than working around them, consistently outperform those that attempt to replace agent judgment. Zoom CX is built around that principle: AI handles the operational load, and agents focus on the interactions that require genuine human connection.
How Zoom CX keeps agents in control while AI handles the heavy lifting
The practical problem Morrissey identifies is what he calls “toggle tax”: the cognitive load placed on agents who must jump between CRM systems, ticketing platforms, billing tools and other applications, mid-call. Each switch costs time and increases error rates, which erodes the quality of the conversation.
Zoom CX’s approach is intelligent context delivery: surfacing only the data an agent needs at the moment they need it, without requiring the agent to navigate to a separate system. When those systems need to be updated at the end of a call, workflow orchestration automates the process. After-call work, currently tracked as a major KPI in most contact center operations, is handled by AI.
“Get rid of that KPI altogether,” Morrissey says.
For complex interactions, AI operates in the background through Expert Assist, providing real-time guidance without interrupting the agent’s conversation with the customer. For routine queries — hours, directions, account status — Zoom CX’s Virtual Agent handles the interaction end to end, delivering precise answers around the clock. Customers choose the experience that fits the problem. Agents step in when the human element is what the moment requires.
AI is here to make human lives better
“AI is not here to replace humans,” Morrissey says. “AI is here to make human lives better. That means making your agents’ lives better and making your customers’ experiences better. If those things aren’t happening, you’ve done something wrong with AI.”
His advice to any CX leader who is six months into an AI deployment and finding the human element getting lost is direct:
- Start with the outcome, not the technology
- Define the experience you are trying to deliver
- Then find the AI that enables it
Many boards expect AI in every interaction as a sign of progress, resisting this pressure is also key.
“Sometimes customers feel pressured to buy AI and try to implement it in unnatural places, and then it doesn’t work.”
To learn more about what Zoom CX can do for your contact center, visit zoom.com/zcx
Frequently Asked Questions: AI-Native Contact Center and Zoom CX
What is an AI-native contact center?
An AI-native contact center is one where AI is embedded into the core platform architecture — not bolted on top of existing infrastructure. It means every component, from virtual agents to human agent assistance to quality management and workforce tools, shares a common data layer. Context flows across interactions automatically. The result is that AI can learn and improve continuously from the full breadth of contact center activity, rather than operating in isolated, disconnected silos.
How does Zoom CX differ from legacy CCaaS vendors?
Legacy CCaaS vendors have typically added AI as a layer on top of existing infrastructure, resulting in disconnected tools that cannot share context. Zoom CX is built on a unified platform where quality management data, workforce management data, virtual agent conversations, and human agent interactions all feed a single AI layer. Handoffs between automated and human agents are seamless, and AI self-improvement is driven by the full history of every interaction across every channel — something a retrofitted legacy system cannot replicate.
What AI features does Zoom Contact Center include?
Zoom Contact Center includes AI Expert Assist for real-time agent guidance, a Virtual Agent for automated customer interactions, CX Insights for actionable business intelligence beyond standard dashboards, resolution dashboards that use AI to evaluate AI performance with human oversight, and workflow orchestration for automating after-call work. These capabilities run on a shared platform, allowing context and learning to flow between channels rather than being isolated within separate point solutions.
Can Zoom CX handle voice, video, and chat in one platform?
Yes. Zoom CX supports voice, video, chat, and messaging from a single agent desktop. Unlike vendors that deliver a unified-looking front end over a fragmented back end, Zoom CX is built on a genuinely integrated platform that shares customer context across every channel in real time. Agents manage all interaction types from one surface, with AI surfacing relevant data based on the live conversation — without requiring manual switching between systems.
What outcomes are Zoom CX customers reporting from AI?
Zoom CX customers are reporting concrete, measurable improvements across resolution rates, wait times, and customer satisfaction. Cricut achieved an 89% reduction in wait times and a 90% reduction in call abandonment after deploying Zoom CX. Vensure Employer Services saw 75% of customer chats handled entirely by the Virtual Agent, with a two-minute average resolution time and 90% positive satisfaction scores. This is formalized in the Agent Performance Suite, announced by Zoom at CCW 2026, which lets organizations test and simulate AI agents before deployment, monitor live performance against resolution. Across deployments, the platform's resolution dashboards — which use AI to evaluate AI performance with humans always in the loop — give customers transparent data on real outcomes, not efficiency proxies.
How does Zoom CX measure and improve AI agent performance?
Zoom CX measures AI agent performance through the Agent Performance Suite, announced at CCW 2026, which applies consistent quality evaluation standards across AI, human, and hybrid interactions. Organisations can test and simulate AI agents before deployment, then monitor live performance against resolution, containment, and CSAT metrics after launch. The suite also identifies successful human-assisted resolutions and feeds them back into the knowledge base automatically, so the system improves with every interaction it handles or assists with.
Is Zoom CX a viable alternative to Genesys, NICE CXone, or Five9?
Yes. Zoom CX competes directly with Genesys, NICE CXone, and Five9 in the enterprise CCaaS market, with a differentiated position built around native platform integration. Where Genesys, NICE CXone, and Five9 have extended their platforms through acquisitions and partnerships, Zoom CX is built on a unified architecture where AI, voice, video, chat, workforce management and quality management share a single data layer which enables a unique level of cross-channel context and AI self-improvement. That same shared data layer powers CX Insights, which lets business and CX leaders query operational and interaction data across Contact Center, Workforce Management, Quality Management, and Virtual Agent in natural language, surfacing patterns and answers that would otherwise require manually pulling and cross-referencing reports from disconnected systems.
Full Interview Transcript: Chris Morrissey at CCW 2026
Click to read the full interview transcript
Rob Scott: Rob Scott here from CX Today, live at CCW Vegas 2026, and I’m here with Chris Morrissey, who heads up the Customer Experience business at Zoom. Chris, how’s it going?
Chris Morrissey: Doing great — nice to meet you, and thanks for being here. It’s been a busy event so far.
Rob Scott: Absolutely. We’re just off the side of the expo floor here, and it’s full of vendors — which leads me straight to my first question. You’re at CCW, surrounded by vendors all claiming to be AI-first. What’s the one thing most of them are getting wrong?
Chris Morrissey: In my opinion, defining themselves as AI-first is the problem. You’re defining yourself by a technology. For our business — and I say this all the time to customers and to our internal staff — we’re a CX-first business. Define yourself by what you’re trying to drive for your customers. We embrace AI, it’s revolutionising the industry, I’m not downplaying that. But we’re not AI-first, we’re ultimately CX-first. If the technology changes, it changes.
Rob Scott: “AI-native” is being used by practically every contact center vendor right now. What does it actually mean to be AI-first or an AI-native solution?
Chris Morrissey: Part of it is a vendor putting itself first. One of the things we need to watch for right now — historically, data silos have been the big challenge. Data spread all over the place. We could now get into a situation where we have AI silos instead. One AI acting as a copilot for agents through Expert Assist, a different AI for your voice virtual agent, a different AI for your chat agents — you’re going to have different experiences across all of those. So I would caution: “AI-native” and “AI-first” are buzz phrases. The question you need to ask is: are you creating a better experience for your customers across every channel? And are you making sure you don’t end up in AI silos?
Rob Scott: How does that compare to building it natively versus bolting it on?
Chris Morrissey: It’s both, honestly — and it’s changing so fast. We went from intent-based AI, creating prompts and then answers, to fully agentic AI remarkably quickly. But what native means for us is that our AI is leveraging all of our QM data, call data, WFM data, and the full history of both human agent and virtual agent conversations. That gives you far more material for self-improvement — making sure both your virtual agents and your human agents get better over time.
Rob Scott: In reality — when a legacy system isn’t truly AI-native and the AI has been bolted on — when does that surface for a CX leader?
Chris Morrissey: One clear example: you have a virtual agent handling calls — and there are a lot of good products out there, I’m not saying they’re good or bad — but at some point a customer needs to reach a human agent. If that human agent has no insight into what happened during the virtual agent interaction, you end up with the agent asking: “Did you try turning it off and on again? Did you try Control-Alt-Delete?” And the customer says: “The virtual agent already told me to do all of that.” By not connecting the journey, you’ve used AI — but you’ve also created friction for the customer. You’ve made the experience worse.
Rob Scott: That makes perfect sense. Customers are being sold a lot of AI promises right now. What are Zoom CX customers actually seeing — and are those outcomes showing up where it matters, on resolution rates and CSAT, or is it still just efficiency gains?
Chris Morrissey: No, it’s showing up in a lot of places. We’re talking to customers every day about what they’re actually achieving. The concept has shifted — it’s not about deflecting calls, it’s about resolving them. Resolving calls with a virtual agent. Helping human agents with guidance and Expert Assist so they can do their jobs better. And then giving the business better insights through tools like CX Insights — tell me something I don’t already know. It used to be reports, then dashboards, and now insights are better than dashboards. It’s helping everyone — the customer, the agent, and the business — to make better, more informed decisions.
Rob Scott: A hot topic at this event — how is Zoom helping customers actually measure AI outcomes and demonstrate AI ROI?
Chris Morrissey: Within our Virtual Agent, we have resolution dashboards. We’re using AI to guard AI — always with a human in the loop — tracking resolution rates, whether they’re improving, and what the human cost of those calls is. That lets us measure the real value of those engagements. Some brands take those savings and bank them. Some reinvest in different models. IKEA is a great example — they saved, I believe, around 8,500 roles using AI, but they didn’t fire 8,500 people. They created new roles. I think you’re going to see a lot more of that. My belief is that the experience customers have with a brand is going to matter more than the product or service that brand offers. That’s where this is heading.
Rob Scott: Most vendors talk about unified platforms — voice, video, chat, messaging, all in one place. But agents are still tabbing between five windows in most contact centers. What does the Zoom agent desktop actually look like day-to-day, and what’s changed?
Chris Morrissey: It’s a great question. And I’d say first — when a lot of vendors say it’s a unified platform, if they’re honest, it’s actually a unified front end. It’s not a unified platform. There’s a lot going on in the background to make it look that way. So make sure you understand what unified really means. Then there’s what we call the “toggle tax” — agents having to jump into a CRM, a ticketing system, a billing system mid-call. Our approach is: based on the context of the conversation, just present agents with the data they need, when they need it — whether that’s from the CRM, the ticketing system, or anywhere else — without making them go to different systems. And not only that: when those systems need to be updated, don’t make the agent do it manually. Use workflow orchestration to automate the after-call work. After-call work is a major KPI right now — all the tasks agents have to complete once a call ends. That should be automated with AI. Remove that KPI altogether.
Rob Scott: When AI is orchestrating all of this guidance in the background, how do you make sure the agent still feels in control — and ultimately that the customer still feels they’re talking to a human?
Chris Morrissey: Sometimes AI is exactly what you want. “What are your hours? What are your directions?” — you get a precise answer 24 hours a day, very quickly. That’s a great customer experience. For complex problems, I want to talk to a human. I want to feel like you value me. That’s when you give human agents the tools to make their job easier. I don’t think AI replaces humans. It lets customers choose — self-serve for the simple stuff, and then when it’s complex, make the agent’s job easier so they can do what they do best: connect with your customers and represent your brand.
Rob Scott: There’s a real tension for CX leaders right now. Boards want AI ROI, customers want to feel heard, and agents are being asked to deliver both at once. Where does Zoom land on that — and what does “human-first, AI-powered” actually mean when the pressure is on?
Chris Morrissey: There’s a lot of pressure coming down from the top to use AI — I agree with that. My advice is: focus on what you’re actually trying to get done. What’s your goal? Be clear, and don’t create AI silos. It doesn’t have to be just about offloading calls. It could be about guiding agents better, getting deeper insights, automating workflows. But once you know what you want to achieve — what problem you’re solving, what experience you’re trying to deliver — then look at the AI products. Don’t start with “we need AI, what are we going to do with it?” Start with the outcome and work back. Sometimes customers feel pressured to buy AI and then try to implement it in unnatural places — and then it doesn’t work.
Rob Scott: That’s the reality for a lot of organisations right now. If a CX leader watching this is six months into an AI deployment and the human element is getting lost — what’s your advice?
Chris Morrissey: At the end of the day, your customers and your agents are all human. They’re all emotional. You can’t take your eye off that. AI is not here to replace humans — AI is here to make human lives better. That means making your agents’ lives better and making your customers’ experiences better. If those things aren’t happening, you’ve done something wrong with AI. It comes back to this: maybe don’t use AI in every single call, every single engagement. AI should make human lives better — not make customers or employees feel like their lives are worse.
Rob Scott: I completely agree. Great advice. Where should someone watching this go if they want to understand what Zoom CX can do for their contact center today?
Chris Morrissey: Head to zoom.com/products/contact-center. Check out our website. We have a huge organisation — and we’re very flat, by the way. So I encourage any customer or partner out there to get in touch directly.
Rob Scott: Great to see you. Thanks so much, Chris.
Chris Morrissey: Thank you.
Rob Scott, CX Today: That’s Chris Morrissey, VP and General Manager of Customer Experience at Zoom. If anything in that conversation got you thinking about where your contact center AI strategy is heading, the link is in the description below. We’ll also have a full follow-up piece on CX Today in the coming days, going deeper on everything we covered here. I’m Rob Scott — thanks for watching, and we’ll see you next time.
