Skip to content
‹ Back to IndustriesMarketing Tech

NiCE's record HMRC deal proves enterprise CX AI is real, but scaling it remains slow

NiCE closed its largest-ever CXone deal with the UK tax authority HMRC. However, the company faces challenges in rapidly scaling AI solutions within the enterprise following contract signing.

This story was produced through MarketScale. See how Marketing Tech teams put it to work with AI Writing.

By MarketScale Newsroom · NiceCxoneCognigyContact Center Ai
Share
Listen to the audio brief

Key facts, context, and what it means.

AUDIO
0:00—
NiCE's record HMRC deal proves enterprise CX AI is real, but scaling it remains slow

Key takeaways

01

NiCE has secured its largest customer experience deal with HMRC.

02

There is a delay between signing AI contracts and scaling its implementation in enterprises.

03

The gap between signing and scaling AI adoption in businesses persists.

Free workspace

Turn your Marketing Tech expertise into content.

Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.

Try it Free

NiCE entered August with its largest contract in company history. A nine-digit total contract value agreement with HM Revenue & Customs, structured alongside Capgemini and Route 101, will see the UK tax authority deploy NiCE's CXone and Cognigy platform to modernize citizen engagement at scale. The deal, reported by both CX Today and CMSWire from NiCE's Q2 2026 earnings call on August 5, also carries an eight-digit annual contract value, making it a record for both the CXone and Cognigy product lines individually.

A second major win also surfaced during the call: an eight-digit ACV agreement with one of the largest healthcare organizations in the United States, again combining CXone and Cognigy, with Accenture handling deployment. Taken together, the two deals make a pointed argument that enterprise-grade CX AI is no longer the territory of early adopters.

Yet NiCE's own executives were careful to frame what winning those deals actually means for the implementation timeline. The gap between signing and scaling, it turns out, is where most enterprise contact center leaders will spend the next 12 to 24 months.

Signing a deal and scaling AI are two very different problems

CEO Scott Russell, speaking on the earnings call as reported by CX Today, described customers as taking a deliberate approach: organizations are working through data preparation, governance structures, and operating model redesign before they attempt to roll AI across every contact center use case. That sequencing is not a sign of hesitation. It reflects the actual complexity of embedding AI into high-volume, regulated service operations.

The contract is the starting line, not the finish. Every major enterprise AI commitment carries a backlog of data readiness and governance work that determines whether deployment takes months or years.

That dynamic showed up in NiCE's financials in a specific way. An analyst on the call noted that net-new AI annual recurring revenue appeared lower in Q2 than in the same quarter a year earlier, despite the company reporting record AI bookings. CFO Beth Gaspich attributed the discrepancy to what she called "the conversion of timing," pointing to the second half of the year for clearer ARR expansion, according to CX Today. For operations and IT leaders evaluating CX AI contracts, the implication is direct: book-of-business growth and realized deployment value are running on different clocks.

NiCE did report that virtually all of its AI revenue in Q2 came from production deployments rather than pilots, and that nearly every enterprise CXone deal during the quarter included an AI component, per CX Today. The company's position is that the market has moved past proof-of-concept. But the ARR timing question suggests that even committed customers are still working through the prerequisites.

Where production AI is already running

NiCE cited two customer deployments as evidence of what operational AI looks like at the far end of that ramp. TripAdvisor, an existing CXone customer, moved from concept to live automated voice calls in two and a half months. According to NiCE's reporting, the AI agent is achieving a 90% customer sentiment score, compared to 71% for human agents handling the same interaction types.

GXBank, Malaysia's first operational digital bank, built its entire CX operation on CXone. NiCE reports that the bank is reaching 95% customer satisfaction and 95% first-contact resolution, with AI autonomously resolving 70% of chat interactions. Both sets of figures come from NiCE directly, and the company did not publish broader averages across its CXone base, so contact center leaders should treat them as illustrative benchmarks rather than typical outcomes.

A 90% AI sentiment score against a 71% human benchmark is the kind of figure that changes a budget conversation, but it takes an unusually clean deployment to get there.

Still, the directional message is clear. In environments where data is well-organized and the use case is well-scoped, AI-led contact center interactions are already outperforming human-only baselines on measurable satisfaction metrics.

Competitive pressure from every direction

NiCE is not building in open space. During the Q2 call, an analyst asked whether the company was running into Genesys, AI-native vendors such as Sierra and Decagon, or Salesforce's Agentforce Contact Center in competitive deals. Russell's answer, as reported by CX Today, was unambiguous: all of the above.

NiCE's stated competitive position rests on two arguments. First, Cognigy is now fully native to CXone, sharing a single application layer, a common data model, and a unified deployment experience rather than operating as a bolt-on. Second, the platform is designed to work across proprietary, open-weight, and future large language models, which Russell framed as a meaningful advantage for organizations that are reluctant to lock into a single AI provider.

That openness argument is increasingly common across enterprise software vendors, but it carries particular weight in contact centers, where the cost of ripping out a core platform is high and the vendor roster around it keeps changing. AI-native specialists can move faster on specific use cases. Incumbents like Salesforce can bundle AI into broader CRM relationships. NiCE's bet is that orchestrating a hybrid workforce across voice, digital, human agents, and AI agents at enterprise scale is a distinct capability that neither category can yet replicate.

Strong results, a stock that did not agree

NiCE beat its own Q2 guidance on both revenue and earnings, according to CMSWire. The HMRC deal alone would have made the quarter newsworthy. But shares fell roughly 6% the morning results were published and were down approximately 4% by the afternoon on the East Coast, per CMSWire's reporting. It was the second consecutive quarter that strong operational numbers failed to move the stock in NiCE's favor, following a similar pattern after Q1 2026 results.

Investor skepticism centered on questions about renewal pricing pressure, the pace of AI ARR growth relative to bookings, and how much of NiCE's AI story reflects its own proprietary technology versus third-party models. Those are fair questions for investors. For contact center and operations leaders, they are less relevant than the underlying operational facts: large enterprises in regulated, high-volume environments are committing to CX AI at the nine-digit level, the deployment ramp is real but manageable when prerequisites are in order, and the competitive field is broad enough that every major vendor is now in the conversation.

NiCE's next ARR test comes in the second half of 2026, when Gaspich indicated the Q2 bookings should begin converting more visibly into recurring revenue. That conversion pace will tell contact center leaders more about enterprise AI deployment timelines than any single landmark deal.

Featured companies

Your experts belong here

Every story in MarketScale Marketing Tech starts with a company putting its practitioners, product marketers, and RevOps leads on the record. Buyers are already reading this topic. The only question is whose experts they find.

Your buyers live in search and AI answers, so published expert content is the channel that compounds instead of expiring.

Book DemoSee how it works15 minutes, straight to a calendar.

About the author

MarketScale Newsroom
MarketScale NewsroomEditorial Team, MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

B2B Weekly

The week in Marketing Tech, and sixteen other industries, every Monday.

Ten stories, one-line takes, five minutes. Free.

Marketing Tech: are you visible to AI?

Before they reach out, Marketing Tech buyers ask AI engines which vendors to trust. Explore how your experts, customers, and partners can become useful content for buyers and AI search.

Free Trial

You just read one Marketing Tech expert. Your company is full of them.

This article was produced through MarketScale. The same platform turns your practitioners, product marketers, and RevOps leads into the articles, video, and social content Marketing Tech buyers are searching for. Start a free trial and see it with your own people. For qualifying companies, no credit card, no demo required.

NPS +73 · 1,000+ creators · 38+ countries

What your free trial includes

Hands-on access to the MarketScale platform
Media requests to your crowd, remote recording, AI writing tools
No demo required. No credit card.
For qualifying companies. Company confirmation required.

More Marketing Tech Insights

Six in ten sales methodology rollouts fade within 90 days

Six in ten sales methodology rollouts fade within 90 days

A Revenue Innovations survey of 419 B2B revenue leaders found six in ten methodology rollouts fade within 90 days and only one in ten last a year. Structured reinforcement showed only a small difference, while the article highlights manager coaching and buyer-side dynamics as key areas to examine ahead of 2027 kickoffs.

  • 01Structured reinforcement teams reported 58% 90-day fade versus 63% with ad hoc reinforcement, suggesting cadence matters less than who does the coaching inside it.
  • 02Respondents estimated only 16% of lost deals went to a competitor, while 44% died inside the buyer’s organization (no decision, failed consensus, or a champion losing influence), suggesting competitor-focused methods may target a minority of losses.
  • 03A 45-day check-in is a signal worth watching for a 2027 kickoff—about halfway to the 90-day mark where most surveyed rollouts had already faded.

Nov 15, 2026

VML study finds 38% of enterprise B2B purchases end with no decision

VML study finds 38% of enterprise B2B purchases end with no decision

VML's Future Buyer Report 2026, a survey of 750 B2B buyers, finds 38% of significant enterprise purchases end with no decision. Generative AI now starts most supplier searches. Finance has the most influence over final approval, though IT leads in telecom, and career worries push 71% of buyers toward familiar suppliers over better-value rivals.

  • 01Defensibility bias is real: 71% of buyers choose a safe, well-known supplier over a rival they believe offers better value, reflecting how approval-stage risk concerns can override perceived value.
  • 02Risk and compliance (39%) and budget or ROI uncertainty (34%) top the stall reasons, suggesting sellers should put ROI, privacy and integration proof in front of finance and risk reviewers early—before late-stage internal alignment drags on.
  • 03A benchmark for any buying committee: an average of 6.8 stakeholders and 5.35 extra weeks of alignment after a preferred supplier has already been picked.

Nov 14, 2026

Cordial’s headless CDP brings CDP queries into ChatGPT, Claude and Copilot

Cordial’s headless CDP brings CDP queries into ChatGPT, Claude and Copilot

Cordial’s latest release positions its CDP as headless: marketers can query it from Claude, ChatGPT, Microsoft Copilot or custom agents via an MCP connector, making the assistant a new front end. For buyers, the focus shifts to what the connector exposes, how fresh the underlying data is, and how MCP access is priced.

  • 01When a CDP answers inside Claude or ChatGPT, the vendor's dashboard becomes optional and the evaluation moves to what the connector exposes and what it costs.
  • 02Assistant access is only as useful as the scores underneath it: layering purchase propensity against discount affinity is what turns a chat question into a full-price send instead of an unnecessary discount.
  • 03Teams whose pipeline lives in a CRM should confirm how that data reaches the assistant, since the systems named at launch are analytics, commerce and data warehouse platforms.

Nov 13, 2026

Explore More Marketing Tech Insights

Read more expert perspectives from across Marketing Tech.

Browse Marketing Tech Hub

About the Expert

MarketScale Newsroom
MarketScale Newsroom

Editorial Team

MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

For B2B teams

Your experts could be publishing here

Stories like this one run on content MarketScale captures from real practitioners. See how your team's expertise becomes coverage in Marketing Tech and beyond.

Book a Demo

Or call us. No forms required. We pick up. 214-945-2512