Under the impact of “SaaSpocalypse,” CIOs and CTOs are taking a tougher stance toward software vendors.

Introduction

In the storm of the “SaaSpocalypse,” enterprise technology executives are taking an unprecedented hard line against software vendors. According to Fortune, since early 2026, shares of Salesforce, SAP, Workday, and ServiceNow have all fallen more than 30%, while the Dow Jones Industrial Average has dropped less than 4% over the same period. The core logic behind this market turmoil is that generative AI and agentic AI are rapidly replicating the functions of traditional SaaS and upending its per-user pricing model. Facing this shift, CIOs and CTOs are no longer taking vendors’ AI roadmaps at face value. Instead, they are redefining the value of enterprise software through shorter contract terms, more prudent dual sourcing, and a more pragmatic outcome orientation.

Event Background: SaaSpocalypse From Risk Consensus to Market Reality

The term “SaaSpocalypse” is not alarmist. When AI startups such as OpenAI and Anthropic began offering AI applications that can replace the functions of CRM, ERP, collaboration tools, and more, the moats of traditional SaaS giants began to crack. The rise of agentic AI has further intensified this pressure—AI agents can autonomously execute tasks, eliminating the need for humans to operate each subsystem individually.

Gartner analyst Arun Chandrasekaran noted that when an AI agent performs tasks, its behavior is often unrelated to humans. If licenses remain tied to humans, that becomes illogical. This view precisely reveals the deep flaw in SaaS per-seat pricing.

Against this backdrop, the mindset of enterprise IT decision-makers is shifting. Allegra Driscoll, CTO of Bread Financial, recalled that in 2023, when communicating with vendors, she asked mostly about AI roadmaps and key milestones; now the conversations go deeper into platform architecture, almost with a philosophical tone. She admitted that in the past she focused more on capacity, security, and data privacy, whereas now she cares about the underlying design of third-party solutions.

Technical Analysis: How AI Agents Break the “Per-Seat” Spell of SaaS

To understand this transformation, one must first understand the fundamental difference between agentic AI and traditional software. Traditional SaaS products, such as Salesforce and Workday, are essentially databases and business process engines that human users operate through an interface. Agentic AI, by contrast, is an intelligent agent that can perceive its environment, make decisions, and execute actions. It can call APIs, operate interfaces, read documents, and communicate with other agents, thereby completing entire workflows.This means that enterprises no longer need to purchase a CRM "seat" for every employee. An AI agent can be configured to handle sales leads, update customer records, send emails, and generate quotes around the clock. Companies only pay for the agent's operational consumption, such as computing resources and API calls, rather than paying on a per-user basis.

This is why the "per-seat" pricing model is being fundamentally challenged. Hadas Reisbaum, CIO of Nice, predicts that as AI agents become widespread, the pricing model of the SaaS industry will undergo significant changes within the next two to three quarters, shifting toward outcome-based pricing. In other words, suppliers will only earn revenue when they deliver measurable business results for customers.

Enterprise Impact Analysis: New Procurement Strategies for CIOs and CTOs

The re-examination of costs is the most direct manifestation of this transformation. Driscoll of Bread Financial says she has set several basic principles: first, not signing agreements longer than one year, because technology changes too quickly; second, for critical use cases, she selects two vendors with similar capabilities simultaneously to observe which one better fulfills its promises—even though this leads to short-term "double spending."

This dual sourcing is not waste, but an investment to reduce the risk of uncertainty. When AI technology is not yet stable, avoiding lock-in to a single vendor is crucial.

From the perspective of capital expenditure (CAPEX) and operating expenditure (OPEX), traditional SaaS models typically classify software costs as OPEX, but AI-agent-driven architectures may change this structure. Enterprises may need to invest upfront in building their own AI orchestration platforms, training specialized models, or procuring GPU computing power, and this portion may become CAPEX. Outcome-based pricing, on the other hand, ties costs to business performance, making IT spending more flexible, but it also requires enterprises to have more refined measurement systems.

At the deployment and operations level, Docusign's CTO Sagnik Nandy provides a very practical decision framework: rank by "dollars, people, and time." First, evaluate the upfront cost of the contract; then confirm how many IT professionals are needed for implementation (he cautions that vendors usually underestimate this number); finally, calculate the time from deployment to measurable value generation. He is especially wary of solutions that improve the efficiency of the CTO team but significantly increase the workload of the CIO.

Security and compliance considerations are also escalating. When AI agents are authorized to access enterprise data, traditional static perimeter security models fail. Enterprises need to ensure that the entire data flow chain is auditable and traceable. Therefore, Driscoll says her conversations with vendors are increasingly like an "architecture review," requiring vendors to explain how the platform is designed, how data flows, and how AI agents are isolated and monitored.From a market performance perspective, the sharp declines in the stock prices of Salesforce, SAP, Workday, and ServiceNow already reflect investor concerns. In contrast, cloud platforms such as AWS, Azure, and Google Cloud may instead become beneficiaries of this restructuring—because no matter which AI agent ultimately wins, they all need underlying computing power and API infrastructure.

AI startups (such as Harvey and Jasper) are riding the wave, but they are not entirely worry-free. Driscoll noted that Bread Financial is building its own agentic AI platform and may eventually replace these AI startups' tools with internal capabilities. This reveals an important trend: companies' long-term goal may be to internalize AI capabilities as their own core assets rather than continuously rely on external vendors.

Ledger's CTO Charles Guillemet proposed two possible future paths: first, large model vendors such as OpenAI and Anthropic invest enormous resources to develop products that seamlessly compete with SaaS functionality, leaving other vendors far behind; second, advances in AI technology will peak at some point, and the competitive focus will shift to optimizing software delivery costs. He personally leans toward the latter and believes that, driven by cost optimization, alternative solutions will become more attractive, but "there is currently no reason to migrate."

Industry Trend Observation: From "AI Roadmap" to "AI-Native Architecture"

Intuit's CTO Alex Balazs shared an interesting phenomenon: in the early days of the AI boom, companies expected that "we build an agent, Salesforce builds an agent, Workday builds an agent, and then these agents communicate with each other." But the reality is that collaboration between heterogeneous agents is far more difficult than imagined.

Therefore, Balazs calls on vendors to "expose tools and skills," which is essentially the AI transformation of the traditional API concept. He calls it "AI API." This means that the future SaaS will not be a closed application, but a set of service interfaces that can be called by AI agents. This evolution will drive SaaS from the "application paradigm" to the "platform and API paradigm."

The long-term impact of this trend is profound. First, enterprise IT architecture will become more modular and orchestrated. Second, data ownership and data interoperability become core concerns. Third, the supplier's value proposition shifts from "providing software" to "providing easily integrable capabilities."In addition, contract durations and business models will also change accordingly. Short-term contracts, pay-as-you-go pricing, and outcome-based incentive clauses will become the norm. For enterprises, this means greater flexibility, but it also places higher demands on internal IT governance capabilities. The multi-cloud strategy will also gain greater momentum—when software can be decoupled into APIs, enterprises are more inclined to distribute different services across multiple cloud platforms to avoid lock-in and optimize costs.

CloudTechDaily Insight

SaaSpocalypse is not a simple market correction, but a structural transformation of the logic of value creation in enterprise software. As AI agents become the core of the enterprise digital workforce, the traditional per-seat SaaS model loses its reason to exist. The winners of this transformation will be those technology platforms that can help enterprises achieve business outcomes at lower cost and faster speed, rather than those organizations that cling to subscription models.

For enterprises, now is the golden window to re-examine the software supply chain. Our recommendations: First, make AI substitutability a mandatory item in software procurement evaluations; second, prioritize suppliers that adopt open APIs and outcome-oriented pricing; third, build internal AI agent orchestration capabilities and treat them as strategic assets on par with cloud infrastructure.

The future enterprise IT architecture will no longer be a "pile of SaaS boxes," but a self-adaptive ecosystem composed of AI agents, APIs, and cloud services. Enterprises that can reshape their own architectures will survive and thrive in this "SaaSpocalypse"; those that still view software as "products" rather than "capabilities" will face the risk of being eliminated.

*This article is analysis based on public reports from Fortune.com and does not represent CloudTechDaily's position on the companies covered.*