CX Providers Set the Pace for Agentic CX This Fall

The Gist

What happened? More than a dozen customer experience technology providers made agentic AI announcements in a single week, spanning trust research, sports partnerships, enterprise case studies and a new category of tooling: agent observability. Why does it matter? Agentic AI is moving from pilot to production faster than most CX teams can build the guardrails to monitor it, and vendors are racing to fill that gap before a bad agent decision costs a brand a customer. What should CX leaders watch? Whether their current stack can trace, evaluate and prove the value of what an AI agent actually did — not just whether it closed the ticket.

Is Agentic Customer Experience getting real? Maybe, if you check out the news lately.

As we close out August today and look ahead toward the final third of 2026, agentic customer experience is showing up in earnings calls, sports sponsorships and product launches all at once.

August was red-hot, and, within the span of a single week, more than a dozen vendors serving the contact center and broader CX stack rolled out news that says this: AI agents are now expected to act on a customer’s behalf, and the market is scrambling to make sure someone — or something — is watching them do it.

The announcements ranged from consumer trust research to a Major League Baseball sponsorship to a brand-new category of tooling called agent observability. Some vendors leaned into what their AI agents can now do autonomously. Others leaned into proving those agents can be trusted, measured and governed.

Here’s what came across my radar lately. It just felt right for a roundup of sorts as we close out the eighth month of the year.

Twilio Says Customers Still Don’t Trust the Bot

Twilio published new research that lands as a reality check for anyone betting the AI agent rollout is going smoothly. According to the company, 78% of consumers have tried to bypass an AI agent to reach a human, and 63% want an easy way to escalate at any point in a conversation — yet only 47% of brands currently offer one.

The company’s blog, authored by Jesse Sumrak, breaks the bypass instinct into three drivers: customers can’t reliably tell when they’re talking to a bot, the stakes feel too high for automation in categories like medical inquiries and payments and — bluntly — the bots still don’t work well enough. Twilio found 80% of consumers report a negative AI experience, and when an agent can’t resolve an issue, it burns an average of 81 seconds before finally escalating.

Twilio’s prescription is to disclose the AI up front, make escalation obvious and make sure conversation history and context survive the handoff to a human agent — a workflow the company says its Conversation Orchestrator and Conversation Memory products are built to support.

Why Do Customers Still Try To Bypass AI Agents?

Twilio’s research points to three drivers: customers can’t reliably tell they’re talking to a bot, the stakes feel too high for automation in categories like payments and medical inquiries, and the bots still don’t resolve issues well enough on their own.

Dialpad Steps Up to the Plate With the Miami Marlins

Dialpad is now the “Proud AI-Powered Communication Partner” of the Miami Marlins, a multi-year deal that puts the company’s Agentic AI platform behind fan communications and in-stadium experiences at loanDepot Park. The pitch is less about answering questions and more about resolving them — a reissued ticket, a moved seat, a real answer delivered before a fan gives up and walks away.

“The bar for fan experience has never been higher,” said Craig Walker, co-founder and CEO of Dialpad, in the announcement. “Fans expect to be known, heard, and actually helped fast, whether they’re calling about a ticket or asking a question at the gate. We’re proud to give them a platform that doesn’t just answer questions — it takes action to fix them, and gets smarter at it every single day.”

Ruben Smith, enterprise technology strategist for the Marlins, framed the deal as a way to modernize communications “without losing the personal connection.” The Marlins join a growing list of sports organizations on Dialpad, including Austin FC, which the company says is already using the same platform to resolve routine fan requests in real time.

What Does Agentic AI Look Like Inside a Stadium?

For Dialpad and the Miami Marlins, it means resolving fan issues — a reissued ticket, a moved seat — before a frustrated fan gives up, extending an approach already running with Austin FC.

Five9 Helps a Ministry Scale Without Losing the Human Touch

Five9 published a customer success story on Christian Healthcare Ministries (CHM), a faith-based health cost-sharing organization that needed to prepare its contact center for a tenfold increase in membership without simply hiring its way through the growth.

CHM’s Wayne Pierce, its CCaaS manager, described the starting point as a 10-option phone menu that routed too many callers into an operator queue, with agents buried in manual work. Working with integration partner Waterfield Tech, CHM moved to the Five9 Intelligent CX Platform and cut its call transfer rate from 48% to 4%. Intelligent routing now identifies roughly 80% of inbound calls before an agent even picks up, and preview dialing cut late-payer collection cycles from 22 days to six — helping drive a 400% increase in collections.

The technology also freed up capacity for something less measurable: outbound welcome calls to every new member, and prayer calls for those navigating difficult health journeys. “Our members just love this,” Pierce said. “They feel loved.” Five9’s Julie LeWarne, who wrote up the case study, notes CHM is now eyeing AI-powered self-service to extend support beyond its current business hours.

Can Agentic AI Scale a Contact Center Without Losing the Human Touch?

Christian Healthcare Ministries cut its call transfer rate from 48% to 4% and used the freed-up capacity for outbound welcome and prayer calls, suggesting automation can create room for more personal contact, not just less.

8×8 AI Studio Interactions Grow Nearly 450%

8×8 reported that customer interactions built on its 8×8 AI Studio have grown nearly 450% since April, with new GDPR compliance and HIPAA-readiness now built into the platform. The headline detail: none of the teams driving that growth are AI specialists. Customers including PrepayPower and LSH Auto describe building working agents in plain language, without an IT project attached.

“Everyone has seen the AI demo,” said Hunter Middleton, chief product officer at 8×8. “Far fewer have seen AI survive contact with production, because between the two sits an integration project most organizations were never staffed to run. AI that demos well was never the hard part. AI that holds up in production, at this pace and across this many industries, is.”

Metrigy analyst Irwin Lazar backed up the shift with research showing nearly 72% of organizations now prefer building custom AI agents — internally or with a partner — over buying an off-the-shelf tool, a preference he attributes to trust and the need for company-specific expertise.

Who Is Actually Building These AI Agents?

According to 8×8 and Metrigy, it’s increasingly not AI specialists — customers describe building working agents in plain language, and nearly 72% of organizations now prefer building custom agents over buying off-the-shelf tools.

View All Verint Puts a Dollar Figure on Empathy

Verint said one unnamed financial institution is estimating $28 million in annual value from real-time, empathy-led coaching rolled out across 6,500 agents. The coaching prompts agents to recognize emotional signals and respond with more consistency during high-stress calls — deployments, family financial needs, government disruptions — and in one high-volume scenario generated real-time guidance on roughly 160,000 calls, along with more than 1 million positive-reinforcement notifications.

“Customer experience improves when agents have the right insights and support at the right time,” said Teresa Anania, chief customer officer at Verint. “Real-time coaching helps large service teams deliver more consistent interactions while contributing to measurable outcomes.”

What Is Real-Time Empathy Coaching Worth in Dollars?

One unnamed financial institution estimates $28 million in annual value from empathy-led coaching across 6,500 agents, with more than 1 million positive-reinforcement notifications generated in a single high-volume scenario.

Google Cloud and Verizon Go All In on Gemini Enterprise

Google Cloud and Verizon announced an expanded strategic partnership built around Gemini Enterprise, which Verizon says now handles the majority of its inbound consumer calls and chats each month. The deal extends beyond customer service into network operations, marketing automation and cloud security, all built on top of Verizon’s multi-year consolidation of legacy data lakes into Google’s Agentic Data Cloud.

“Serving each and every one of our customers by name requires working AI-first at every level,” said Alfonso Villanueva, Verizon’s chief transformation officer and EVP of Verizon Consumer. Google Cloud’s Karthik Narain, chief product and business officer, called Verizon’s approach a template for “what a true, full-scale AI transformation looks like for a global enterprise.”

What Does Full-Scale Enterprise AI Transformation Actually Look Like?

For Verizon, it means Gemini Enterprise now handling the majority of monthly consumer calls and chats, extending beyond customer service into network operations, marketing automation and cloud security.

Operata Launches the Control Layer CX Has Been Missing

Operata introduced expanded CX AI Agent Observability, positioning itself as the control layer that sits outside the fragmented stack of telephony, CCaaS, CRM and AI platforms most enterprises now run — normalizing metrics, traces and logs from every service into a single record of the customer journey.

“At Operata, we believe human judgment and AI execution together deliver the best end-to-end customer experience,” said Romilly Blackburn, co-founder and CEO of Operata. “Observability is the control layer for both; it monitors, assures, diagnoses, and optimizes them as they work side by side.”

The timing isn’t incidental. Article 50 of the EU AI Act, which requires disclosure for any AI system a person interacts with, has applied since August, and Operata frames its platform as the evidentiary record enterprises will need to prove a handoff carried context, or that a required disclosure was actually audible to the customer rather than just technically played.

Is Agentic CX Actually Ready for Production?

The vendor activity suggests yes — but only alongside new observability, disclosure and escalation tooling built to catch what autonomous agents get wrong before customers do.

Zendesk Gives Agents a Better Memory

Zendesk quietly shipped a smaller but consequential update: the ability to search for related custom object records — orders, contracts, returns, entitlements — directly inside action flows, without already knowing the record ID. It’s the kind of plumbing update that doesn’t make headlines but determines whether an AI agent can actually pull the context it needs to resolve something, or whether it stalls out waiting on a human to look something up manually.

Why Does a Small Search Update Matter for AI Agents?

Zendesk’s update lets action flows search for related records like orders and contracts without already knowing the ID — the kind of plumbing that decides whether an agent resolves a case or stalls out waiting on a human.

Omilia Bets on a Five9 Veteran to Crack DACH

Omilia named Michael Weiss, a four-year Five9 veteran, as vice president for Germany, Austria and Switzerland, betting his enterprise sales background can move the region’s notoriously cautious banks and insurers off legacy IVR and first-generation automation.

“German, Austrian, and Swiss organizations are among the most cautious adopters of contact center AI, and for good reason, given the high compliance bar in banking and insurance,” Weiss said. “The caution isn’t reluctance, it’s a need for proof of ROI and sovereign cloud.” Omilia’s chief revenue officer, Nick Delis, called DACH “one of the most demanding enterprise markets in the world, with exacting compliance requirements and long buying cycles, but once you’re in, you’re in for the long term.”

Why Is the DACH Region So Resistant to Contact Center AI?

Omilia’s new regional VP frames it as caution rather than reluctance — German, Austrian and Swiss banks and insurers face a high compliance bar and want proof of ROI and sovereign cloud before moving off legacy IVR.

IBM and the USTA Level Up the US Open

IBM and the United States Tennis Association (USTA) rolled out new AI-powered fan features for the 2026 US Open, including a Serve Quality metric that tracks 21 data points across a player’s body and racquet 50 times per second and an enhanced Match Chat conversational assistant that now returns photos and video alongside text answers.

“Our work with IBM this year has been game-changing, pushing the boundaries of what’s possible in digital fan engagement,” said Brian Ryerson, senior director of digital strategy at the USTA. IBM’s Jonathan Adashek, senior vice president and chief global affairs officer, said the collaboration lets IBM “take millions of live data points streaming from the court and automatically route them into workflows that deliver usable insights in minutes and hours.”

How Is AI Changing the Live Sports Fan Experience?

IBM and the USTA’s 2026 US Open rollout includes a Serve Quality metric tracking 21 data points per player 50 times a second, plus a Match Chat assistant that now returns photos and video alongside text.

BrightEdge Says Your Website Isn’t Ready for AI Agents

BrightEdge launched Agent Edge, a platform built around what it’s calling the “Second Web” — a growing layer of AI agents researching, comparing and taking action on a person’s behalf, often before that person ever visits a brand’s website. The number that should worry every marketer: BrightEdge found 97% of the top 20,000 websites aren’t currently set up to serve AI agents in optimized formats, forcing agents to burn tokens parsing HTML and navigation elements built for humans, not machines.

“AI agents are getting dramatically better at doing the work people once had to do themselves,” said Jim Yu, CEO and co-founder of BrightEdge. “The companies that can understand and optimize those interactions will have a significant advantage as this second audience grows.” In an early deployment with chip designer Arm, BrightEdge says it cut the payload served to AI agents by roughly 90%, more than doubling verified agent traffic within weeks. Bloomfire’s Brian Zander, an early Agent Edge customer, reported a 30% lift in direct AI referral traffic. The move lands squarely in the broader generative engine optimization conversation reshaping how brands get found.

Vendor Moves at a Glance

Editor’s note: The following table summarizes the agentic CX and agent observability announcements covered above, based on vendor press materials and public statements from late August 2026.

VendorWhat They AnnouncedWhy It MattersTwilioResearch showing 78% of consumers try to bypass AI agentsQuantifies the trust gap brands need to close before scaling automation furtherDialpadMulti-year Agentic AI partnership with the Miami MarlinsSignals agentic CX moving into high-visibility consumer brand experiencesFive9Case study on Christian Healthcare Ministries’ contact center transformationShows measurable ROI — a 44-point drop in call transfer rate and 400% collections growth8x8450% growth in AI Studio interactions since April, plus GDPR/HIPAA readinessReinforces demand for no-code, natural-language agent buildingVerint$28 million in estimated annual value from empathy-led agent coachingPuts a concrete financial figure behind soft-skill coaching investmentsGoogle Cloud & VerizonExpanded Gemini Enterprise partnership across CX, network and marketingOne of the largest enterprise AI deployments in telecom to dateOperataExpanded CX AI Agent Observability platformEstablishes a single record of AI and human interactions amid rising AI disclosure regulationZendeskCustom object record search inside action flowsRemoves a common blocker to AI agents resolving tickets independentlyOmiliaNew VP hire to lead DACH region expansionTargets one of the most compliance-heavy, automation-resistant enterprise marketsIBM & USTANew AI-powered fan features for the 2026 US OpenDemonstrates agentic AI applied to real-time, high-volume consumer engagementBrightEdgeLaunch of Agent Edge for AI-agent-optimized web contentFinds 97% of top websites unprepared for AI agent traffic

CX Providers Set the Pace for Agentic CX This Fall

Tags: