NiCE has plenty to celebrate after its second-quarter earnings call, but the vendor also revealed the work customers must do before AI can deliver at scale.
The company secured its largest-ever CXone and Cognigy deal, a nine-digit total contract value agreement with HM Revenue & Customs. It also landed another eight-digit annual contract value win with one of the largest healthcare organizations in the US.
But behind the landmark deals and upbeat AI adoption figures, NiCE also offered a more grounded assessment of where enterprise customer experience AI is today.
Customers are buying in, but they are not necessarily ready to scale straight away, as Scott Russell, CEO of NiCE, explained:
“Customers are taking a measured approach as they prepare their data, the governance, [and] the operating models before they scale AI across all the use cases."
This detail is particularly significant for contact center leaders. The industry’s agentic AI conversation is increasingly focused on deployments, not pilots. Yet NiCE’s call suggests that getting from a signed contract to a fully operational AI-enabled CX operation remains a slower, more complicated process.
NiCE Lands Its Largest-Ever CXone Deal
The headline from the call was undoubtedly the HMRC win.
Working with Capgemini, NiCE will provide its unified CX AI platform to help the UK tax authority modernize citizen engagement.
NiCE described the agreement as an eight-digit ACV deal and a nine-digit TCV deal, its largest ever for both CXone and Cognigy.
That is a major statement of intent from a public-sector organization with a vast, highly regulated service operation. It also offers a useful counterpoint to the idea that agentic AI is largely limited to low-risk customer service use cases.
Meanwhile, NiCE said it had also won an eight-digit ACV agreement with a large US healthcare organization. The customer selected CXone and Cognigy, with Accenture involved in the deployment, to advance customer engagement on a unified AI platform.
Production AI Is Growing, But the Conversion Gap Remains
NiCE was also keen to stress that it is seeing real AI deployments, rather than a collection of proof-of-concepts.
Russell said that “virtually all” of the company’s AI revenue is now coming from production deployments. He also stated that nearly every enterprise CXone deal during the quarter included AI.
There are examples to support the claim:
TripAdvisor, an existing CXone customer, reportedly moved from concept to its first automated voice calls in two and a half months. NiCE said the company’s AI agent has achieved a 90% customer sentiment score, compared to 71% for human agents.
GXBank, Malaysia’s first operational digital bank, has built its CX operation on CXone. According to NiCE, the bank is achieving 95% customer satisfaction and 95% first-contact resolution, while AI autonomously resolves 70% of chat interactions.
Those are eye-catching figures. However, they are vendor-reported customer examples, and NiCE did not provide broader data on average containment rates, implementation timelines, escalation levels, or the number of CXone customers now running agentic AI at scale.
The Q&A also exposed a timing issue.