From feature wars to outcome battles
SaaS business model transformation in the AI era means software vendors can no longer win by piling on features; they now have to prove they can deliver measurable business outcomes, build durable operating capabilities, and fit into complex enterprise environments where agents and humans work together across many systems. AI is making some parts of software easier to emulate, making generic functionality more abundant, exposing how little differentiation many vendors had beyond a polished interface, roadmaps and sales stories. In mature categories, buyers assume every credible product hits a functional baseline: a CRM manages customer data, a DAM stores and finds assets, a workflow tool routes work — and most of them do. The competition has moved away from feature parity to a harsher test: which SaaS vendor can help a customer turn those features into a working, dependable operating model rather than an expensive configuration project that people ignore.
AI software commoditization and the disintermediation shock
Agentic AI is not killing SaaS, but it is stripping away its old comforts. When you can say, “I need these things,” and a model like Claude, ChatGPT or Perplexity writes code and assembles capabilities on demand, that ephemeral application becomes the new work surface and you are disintermediated from the underlying app. According to technology analyst Gartner, up to USD 234 billion (approx. RM1,076 billion) in application spending is exposed to agentic arbitrage between now and 2030, where agents complete tasks across multiple systems instead of users clicking through interfaces. By the end of the decade, AI interactions will account for roughly 20% of enterprise application SaaS spending. Markets have already wiped about USD 300 billion (approx. RM1,380 billion) off SaaS valuations in 18 months on fears that agents will replace traditional tools. Yet this “SaaS apocalypse” is overrated: the real story is disintermediation and metamorphosis, not extinction.
Enterprise software strategy: selling capability systems, not logins
If AI makes features cheap, the only defensible strategy is to sell complete capability systems. Enterprise technology does not create value as a product; it creates value as part of a system of connected decisions. The software product market is becoming a market for capability systems that combine tools with data, integrations, permissions, workflows, governance, skills, operating roles, partner models and clear value measures. Two organizations can buy the same platform and get opposite results: one boosts output quickly, the other spends 18 months configuring workflows that nobody follows, struggles to migrate content, loses trust in data and falls back to manual workarounds. The more durable moat is operational consequence: how deeply a platform embeds into the operational, financial, governance or decision fabric of the business. Leading vendors will be those that give customers the most credible path from product purchase to dependable operating capability, not the longest feature list or flashiest demo.
Workday and the new playbook for SaaS vendor differentiation
Major SaaS players are already rewiring their position in response to AI disintermediation impact. One executive at Workday argues that in a period of AI uncertainty, buyers will favor providers that can offer certainty in outcomes. Workday is investing in becoming the “front door to work,” where employees log in once, ask natural-language questions about issues such as payroll variations, and receive tailored answers from enterprise data sources through agentic services. This is not a cosmetic AI add-on; it is a bet that durable capabilities are what you sell, and an app that is only logic and functions is now a weak proposition. More broadly, founders are no longer selling access to a tool; they are selling units of work. Domain expertise becomes the moat in a world where anyone can generate code, pushing SaaS vendors to go vertical and tie their systems tightly to industry-specific workflows and decisions.
From SaaS to Service-as-Software: where the value is moving next
The shift under way is not a downturn; it is a repricing of what software is for. Markets are moving value from tools that help a human work to software that does the work itself. Foundation Capital estimates the opportunity for “Service-as-Software” at around €4.02 trillion (approx. RM20.2 trillion) or USD 4.6 trillion (approx. RM21.2 trillion), compared with a global SaaS market near €175 billion (approx. RM880 billion) or USD 200 billion (approx. RM920 billion). Per-seat pricing assumed a human in the seat, but when an agent does the work of ten people, ten seats do not follow, and pricing models are shifting toward consumption. Roughly 77% of the largest software companies already use some form of consumption pricing. When software sold seats, it was paid from the IT budget; when it does the work, it competes for the labour budget, which is much larger. None of this will be smooth: more than 40% of agentic AI projects may be cancelled by 2027 due to cost, weak governance or unclear value. But “SaaS is dead” is the wrong headline; the thing called SaaS is being reborn as outcome-centric service software.
