Finvest
NICE Software · Cloud software · AI · Customer service · Thesis updated July 17, 2026

AI demand is real, margins are the test

01 Running thesis

AI backlog versus margin pressure

The bull case is that NICE is catching the right AI wave. Its AI ARR, or annual recurring revenue tied to AI products, rose 66% year over year to $328 million. Cloud backlog grew 25% year over year, or 22% organically, which gives management a long runway of contracted work to turn into revenue.

Management expects cloud revenue growth to reaccelerate to 14.5% to 15% in 2026. That matters because cloud is already 77% of total revenue. If large enterprise deals keep going live, the growth story can improve without NICE needing to rebuild the whole business.

The bear case is that the company must spend more to win this AI cycle. Q4 gross margin was 69.3%, and management plans higher investment in 1H 2026 across cloud costs, R&D, and sales. Cloud net revenue retention, which measures growth from existing cloud customers after churn, has also settled at 109%, below the stronger levels investors used to expect.

The tension is clear. NICE may be using AI to protect and expand its contact-center platform, but AI could also change how customers buy it. The key question is whether consumption pricing adds a new growth layer, or whether it slowly replaces old seat-based licenses tied to human agents.

Feb 2026The 2025 20-F added clearer AI risk language. NICE now explicitly flags the risk that AI vendors train on its outputs and that demand may fall for products priced by human-agent seats.
Feb 2026Q4 showed stronger AI momentum, with AI ARR up 66% year over year to $328 million. Management also guided 2026 cloud growth to 14.5% to 15%, but warned that 1H 2026 margins will absorb higher investment.
Nov 2025Q3 confirmed that Cognigy closed early and AI ARR kept rising, but cloud NRR slipped to 109%. The thesis stayed balanced between strong AI demand and weaker expansion signals.
Aug 2025Q2 showed AI and self-service ARR rising 42% year over year to $238 million, but LiveVox churn was worse than expected. Gross margin also dipped to 69.3% as NICE invested in international cloud infrastructure.
May 2025Q1 added useful disclosure, including cloud NRR of 111% and AI ARR above $200 million. Large enterprise wins supported the long-term view, even though revenue from mega-deals was not expected until 2026.
Mar 2025The 20-F confirmed Scott Russell became CEO on January 1, 2025. The cloud, AI, and product thesis did not materially change.
Feb 2025Q4 2024 results were solid, but 2025 cloud growth guidance moved down to 12%. The update confirmed that complex AI enterprise deals were taking longer to deploy and recognize as revenue.
Nov 2024Q3 2024 showed strong cloud ARR and fast Copilot adoption, but also longer implementation timelines. The long-term AI case improved while the near-term revenue timing got harder.
02 Business model

Recurring cloud, rising AI costs

NICE makes most of its money from cloud software subscriptions. Customers use its CXone platform to route service calls and chats, manage agents, automate tasks, and analyze customer interactions. The same company also sells Actimize, a financial crime and compliance product used by banks and other regulated firms.

The model is sticky because big contact-center systems are hard to replace once they are running. Large enterprises often take a long time to implement the software, but once live, they can stay for years. That is why cloud backlog and net revenue retention matter so much.

AI is changing the pricing mix. NICE is moving more toward interaction and consumption-based pricing, where customers pay based on usage instead of only by the number of human agents. This can help if AI creates more billable work, but it can hurt if fewer human-agent seats are needed over time.

The main break point is cost. AI workloads need more cloud infrastructure, more product work, and more sales help. If those costs keep rising faster than revenue, NICE could grow while still producing weaker margins.

03 Product portfolio

The stack NICE is building

Cash cow

CXone

CXone is the core cloud platform for customer experience. It handles the contact-center work that ties customers, agents, managers, and workflows together.

Growth engine

CXone Mpower

CXone Mpower is the main AI-led platform push. NICE is using it to sell end-to-end customer service automation to large enterprises.

Growth engine

Copilot and Autopilot

Copilot helps human agents work faster, while Autopilot handles more automated service tasks. These products are central to the shift from old contact-center software to agentic AI.

Growth engine

Cognigy

Cognigy adds conversational AI and virtual-agent tools. NICE closed the acquisition in September 2025 and plans deeper native CXone integration in late 2026.

Steady

LiveVox

LiveVox adds outbound contact-center capabilities. The business had unexpected churn in 2025, but management later said it was stabilizing.

Steady

Actimize

Actimize serves financial crime and compliance customers. It is smaller than Customer Engagement but gives NICE a second software line outside contact centers.

04 Business segments

Customer engagement dominates

Customer Engagement84%modest
Financial Crime and Compliance16%modest

The segment mix is from Q4 2025 revenue. Customer Engagement was 84% of revenue, while Financial Crime and Compliance was 16%, so results are still heavily tied to contact-center spending.

05 Risk factors

What could go wrong

AI costs outrun revenue

High impact · Medium odds

NICE is deliberately spending more in 1H 2026 on cloud infrastructure, R&D, and sales to scale AI. Q4 gross margin was 69.3%, so there is not much room for investors to ignore cost pressure. If AI usage grows but compute and support costs rise faster, profit quality could weaken.

We watchWatch gross margin after 1H 2026 and whether management says the investment peak has passed.

Seat pricing gets cannibalized

High impact · Medium odds

NICE has long sold software tied in part to human agents. Management's 20-F now calls out the risk of lower demand for products priced by the number of human agents deployed. If customers use AI to reduce seats, consumption revenue must grow enough to fill the gap.

We watchWatch cloud net revenue retention, AI ARR growth, and any comments on agent-seat volumes.

AI-native rivals copy the workflow

Medium impact · Medium odds

The 20-F added a specific risk that third-party AI vendors may train models on NICE inputs, outputs, or product behavior. That could help rivals copy or approximate parts of NICE's system. The risk is bigger if customers start buying smaller AI tools instead of a full platform.

We watchWatch win rates against AI point solutions and management comments on IP protection.

Large deals take too long to go live

Medium impact · Medium odds

NICE is winning large enterprise projects, but those deployments can take time. That delays revenue even when bookings look strong. Cloud backlog grew 25% year over year, which is good, but investors still need to see that backlog turn into live revenue.

We watchWatch cloud revenue growth versus backlog growth and updates on large enterprise go-lives.

Existing customer expansion stays muted

Medium impact · Medium odds

Cloud net revenue retention was 109% for the trailing 12 months of Q3 2025. That is still positive, but lower than the kind of expansion investors want from a leading cloud software platform. If NRR stays near this level, NICE needs more new sales to hit its growth goals.

We watchWatch whether cloud NRR moves back above 109% as Cognigy and Mpower are integrated.
06 Quick answers

In one breath

What does NICE Ltd. actually do?

NICE sells cloud software that helps companies run customer service across calls, chats, and digital channels. It also sells financial crime and compliance software through Actimize.

Why is AI so important to NICE?

AI is becoming a larger part of how customer service work gets done. NICE says AI ARR reached $328 million and grew 66% year over year, which makes AI one of the clearest growth engines in the business.

What is the biggest risk for NICE stock?

The biggest risk is that AI changes the contact-center model faster than NICE can profit from it. If customers need fewer human-agent seats, NICE must replace that revenue with usage-based AI revenue while keeping margins under control.