Has the “SaaSpocalypse” Come for CX Technology Providers?
Through 2025 and into this year, a growing number of investors and commentators have argued that software providers are entering a “SaaSpocalypse.” The basic assertion is that AI is starting to break the economics that supported software valuations for the last decade, leading to bearish investor sentiment. Five claims underpin this view. In brief, AI may:
- Push providers away from predictable subscription pricing toward less familiar usage- or outcome-based models.
- Reduce the number of users companies need.
- Make it easier for customers to build tools themselves.
- Lower barriers for new competitors.
- Increase SaaS providers’ operating costs.
This note tests these five claims, focusing specifically on customer experience (CX) technology. CX is a pertinent market to examine because it sits close to where AI is already being deployed at scale: marketing, sales, commerce, customer service, contact centers, customer engagement, and retention. In other words, if AI were already causing major disruption to software demand, pricing, or profitability, CX would be the canary in the coalmine. There are also many publicly traded providers operating in CX (including the hyperscalers and platform providers), enabling broad analysis of these companies’ most recent full-year results.
This note’s analysis finds that the SaaSpocalypse narrative is largely false for CX markets. The latest full-year results reveal that most CX vendors still grew, many preserved or improved profitability, and several of them disclosed rising AI-related ARR, backlog, bookings, or usage.
However, there are core initiatives CX vendors should be focused on through 2026’s fiscal year to avoid turning investor anxiety into self-reinforcing commercial pressure. Falling software valuations may not be indications that sector fundamentals are breaking down, but they do reflect investor uncertainty about AI costs, disclosure complexity, and changing monetization models. This note thus concludes by exploring what those core initiatives could look like and how CX technology buyers and vendors alike should be navigating this market.
Methodology
This note analyzes the latest available full-year results for 24 public companies with significant CX offerings. The companies are 8x8, Adobe, AWS (i.e. Amazon Connect), Braze, Cisco (i.e. Webex Suite), Five9, Freshworks, Google (i.e. Customer Engagement Suite), HubSpot, Klaviyo, Microsoft (i.e. Dynamics 365 CRM), NiCE, Oracle (i.e. Oracle Cloud CX), Pega, RingCentral, Salesforce, SAP (i.e. SAP Customer Experience), ServiceNow (i.e. ServiceNow Customer Service Management), Shopify, Sprinklr, Sprout Social, Twilio, Verint (before their acquisition by Thoma Bravo in November 2025), and Zoom (i.e. Business Services).
For Microsoft, Oracle, SAP, Google, AWS, and Cisco, the read-through is necessarily directional because those companies do not report a standalone CX P&L; CX sits inside broader segments such as productivity, cloud, collaboration, or enterprise applications.
Claim 1: The traditional seat-based SaaS model in CX is being disrupted by usage-, automation-, and outcome-based pricing, creating reporting noise and valuation compression even where demand remains healthy.
Result: Strongly supported.
Of the five claims, this is the one most clearly supported by the evidence. Verint explicitly said revenue and non-GAAP EPS were affected by unbundled SaaS revenue even as ARR was ahead of guidance and new-deal SaaS ACV bookings rose 30%. RingCentral said ARR from customers using at least one monetized AI product more than doubled and was approaching 10% of total ARR, while its new AI-led products reached $100 million of ARR. Salesforce increasingly frames monetization through Agentforce and Data 360 ARR; Amazon Connect reached a $1 billion revenue run-rate on a usage-driven model; and 8x8 highlighted accelerating usage-based revenue from AI offerings. Twilio, Braze, and Klaviyo reinforce the point because their models are already centered more on engagement, data, and usage intensity than on classic seat counts.
This is the best explanation of why market sentiment can look worse than operating performance. Revenue is still there, but the unit of monetization is changing. Investors are trying to value these businesses with old heuristics (seat growth, clean ARR comparability, and familiar margin ladders), while vendors are moving toward usage, automation, and outcome-linked economics. This is why reported revenue can temporarily look weaker even when underlying commercial momentum is healthy: The monetization model is changing faster than investors’ comparability frameworks.
Claim 2: AI automation is reducing the number of licensed users, agents, and support staff required to run CX operations, causing seat compression and weakening recurring revenue growth.
Result: Not supported.
The full-year results do not support this claim as a general sector explanation. Microsoft 365 Commercial still posted 6% seat growth. HubSpot increased customers by 16% and increased average subscription revenue per customer by 3%. ServiceNow reported substantial growth in licensed users, workflows, and transactions, while current remaining performance obligation (cRPO) grew 25%. Zoom’s enterprise revenue grew 6.5%, and its $100,000-plus customer cohort grew 9.3%. Freshworks improved net dollar retention to 108%. Twilio ended the year with more than 402,000 active customer accounts and 108% full-year dollar-based net expansion rate (DBNER). And Braze reported 109% dollar-based net retention (DBNR). These are not the signatures of widespread seat-led demand failures.
The absence of broad seat compression does not mean AI is having no effect. It means that seat count is becoming a weaker standalone indicator because vendors are monetizing more through usage, automation, and outcomes. The set of indicators that will become more common include consumption or usage-based revenue as a percentage of total revenue (demonstrating successful pricing model transitions), gross margin trajectory or net AI compute costs (showing whether AI features are adding value or diluting it), remaining performance obligation growth, dollar-based net retention, and platform revenue concentration.
Claim 3: Generative AI development tools have reduced the cost of custom software enough that enterprises are substituting self-built CX capabilities for commercial platforms, impairing growth for packaged CX vendors.
Result: Not supported at enterprise scale.
The full-year results do not show that this claim is happening at scale in enterprise CX. For Salesforce, Agentforce and Data 360 ARR exceeded $2.9 billion (including Agentforce ARR of $800 million and $1.1 billion in Informatica Cloud ARR following the completion of that acquisition). Implied organic Agentforce and Data 360 ARR was approximately $1.8 billion, up over 100% year over year, with more than 60% of Q4 bookings coming from existing customer expansion. ServiceNow said Now Assist net new ACV more than doubled year over year. Pega increased ACV by 17% and Pega Cloud specifically by ACV 33%. NiCE increased AI ARR by 66% and said AI was included in 100% of new seven-figure CXone deals in 2025. Freshworks said Freddy AI surpassed $25 million of ARR, while Braze, Twilio, and Klaviyo all still produced strong growth at 24.4%, 14%, and 32%, respectively.
The likely reason for this growth is architectural. Enterprises can build lightweight automations, assistants, and edge experiences more cheaply than before, but the core CX problem still includes data integration, workflow, channel orchestration, governance, analytics, compliance, and operational reliability. Those remain platform problems. “Build” is becoming more viable for extensions; it is not yet displacing “buy” for the enterprise core.
Claim 4: AI has lowered the technical and capital barriers to launching CX products, increasing competition and weakening incumbent growth, retention, and pricing power.
Result: Partially supported, but the effect is highly uneven.
The full-year results support a narrower version of this thesis, not a universal one. Sprinklr is the clearest pressure case: Full-year subscription revenue grew only 5%, RPO was flat, cRPO grew 1%, and gross margin fell meaningfully. Sprout Social also showed some competitive strain, with 2025 dollar-based net retention at 100% (down from 104%) and 102% excluding SMB (down from 108%). Mature communications vendors were slower too: RingCentral grew 5%, Zoom 4.4%, and 8x8’s progress was positive but still transitional. These data points are consistent with greater feature-level competition and tighter expansion dynamics in selected categories.
Yet, the counterexamples are too strong to support a general collapse. Braze grew 24.4%, Klaviyo 32%, Freshworks 16%, HubSpot 19%, Adobe Digital Experience 9% with 11% subscription growth, Shopify 30%, and Five9 reported record full-year revenue of $1.1 billion with Q4 enterprise AI revenue growth of 50%. That pattern suggests barriers have fallen most for solutions with niche features, not for platforms with broad data consumption, an installed-base distributed across the organization, or workflow depth across modules.
Claim 5: Rising AI inference, model-serving, and data-processing costs are structurally weakening the unit economics of CX software and driving margin erosion across the sector.
Result: Partially supported, but concentrated at the infrastructure layer.
The most visible cost pressure is showing up in the hyperscalers and cloud foundations, not as a generalized collapse in application-layer CX margins. Microsoft said gross margin percentage in Productivity and Business Processes decreased slightly because of scaling AI infrastructure. Amazon’s trailing 12-month free cash flow fell sharply because purchases of property and equipment increased by $50.7 billion, primarily reflecting AI investments (even as AWS sales rose 20% to $128.7 billion and AWS operating income rose to $45.6 billion). Alphabet finished 2025 with Google Cloud at an annualized revenue run rate above $70 billion and signaled very large 2026 CapEx, while Oracle’s Q4 cloud revenue grew 27% but non-GAAP operating income grew only 5% ̶ consistent with heavier cloud investment.
The application layer picture is much healthier. Salesforce reported strong non-GAAP operating profitability, while HubSpot, Freshworks, RingCentral, and Zoom also delivered healthy margins. Twilio generated $945.4 million of free cash flow, and Shopify produced a 17% free-cash-flow margin. The anomalies are selective rather than systemic: Sprinklr showed margin pressure, and Braze and Klaviyo both showed modest gross-margin pressure while still improving operating leverage and cash generation. The trend, therefore, is upstream compute-cost pressure combined with downstream resilience at the application layer.
Overall Themes
Taken together, the evidence does not demonstrate a generalized “SaaSpocalypse” in CX. It points instead to three narrower themes. First, AI cost pressure is concentrated upstream in infrastructure and hyperscaler capital expenditure. Second, lower barriers to entry matter mainly in thinner, more replaceable feature areas. Third, the biggest shift is the transition away from purely seat-based SaaS economics toward hybrid models based on usage, automation, and outcomes. In other words, this is better understood as a monetization and disclosure reset than as a demand collapse.
That also helps explain why investor sentiment has looked harsher than the operating results. Several diversified vendors do not report CX separately, which makes it easy for the market to overgeneralize from broader software or cloud narratives. At the same time, AI monetization is often showing up first in ARR, bookings, usage, backlog, and new product metrics rather than in a neat step-change in reported GAAP revenue. The market is therefore also reacting to disclosure complexity and valuation-model mismatch.
How CIOs and tech buyers should navigate this environment
CIOs and technology buyers should behave as though CX is in a commercial and architectural transition. That means continuing to modernize core CX capabilities but being more selective about what is considered strategic. The operating evidence still favors durable platforms: Salesforce, ServiceNow, Pega, NiCE, HubSpot, Freshworks, and Adobe all showed either strong growth, healthy margins, meaningful AI solutions, or a combination of the three. The right response is to consolidate around platforms with proven data, workflow, and cross-channel depth while keeping narrower use cases more modular.
Moreover, as pricing models shift, buyers should insist on terms that can survive the shift from seats to usage and outcomes: clear definitions of billable units, caps on AI overages, auditability of usage, rebasing rights, credits where automation reduces seat needs, and strong data portability and exit rights. Vendors are moving toward new monetization units, from Agentforce-style AI ARR to monetized AI products and usage-based contact center models.
On architecture, the best posture is usually “buy the core, build the edge.” Buyers can increasingly use AI development tools for light extensions, task-specific agents, and differentiated workflow logic. But the evidence does not support building the entire CX stack in-house when the heavier problems are still orchestration, governance, compliance, analytics, and systems integration.
How vendors themselves should navigate this environment
Vendors need to make the market’s job easier, and the first step is better disclosure. Where possible, they should break out AI ARR; attach rates; retention by customer cohort, backlog, or RPO; and the mix between seat-based and usage-based revenue. The more diversified vendors have a particular challenge here because the market cannot easily isolate CX performance inside broader segments. The more focused vendors already providing clearer AI and retention signals are generally giving investors a better framework for interpretation.
Second, vendors need to prove AI economics rather than AI activity. The market is no longer short of product announcements; it wants evidence that AI improves growth quality, expansion, retention, margin, and cash generation. Freshworks’ Freddy AI ARR, NiCE’s 66% AI ARR growth, Salesforce’s Agentforce and Data 360 ARR, Twilio’s strong cash generation, Braze’s growth-plus-profitability profile, and Klaviyo’s combination of 32% growth with strong free-cash-flow generation are all the kinds of metrics that shift the discussion from narrative to fundamentals.
Third, vendors should tell a platform story, not a feature story. The most defensible platforms are the ones that look like systems of engagement and execution (combining data, channels, workflow, automation, and governance) – not the ones selling isolated AI features. They also need to reduce investor confusion during pricing model transitions by providing bridge metrics that reconcile older seat-based views with newer usage- and outcome-based economics. Verint shows why that matters: a company can be commercially healthier than its reported revenue trend initially suggests when the monetization model changes faster than the market’s frame of reference.