SaaS was meant to free us from the shackles of legacy IT, promising agility, scalability, and simplicity. Enterprises rushed to adopt cloud-based applications, dumping costly custom code for always-on access to ‘best-practice’ business logic. But we didn’t get agility. We got uniformity. Services-as-Software – powered by on-the-fly AI gives us the opportunity to reclaim individuality and, with it, build responsive organizations that are better able to meet customer needs.
SaaS forced enterprises into cookie-cutter workflows, embedding standard and repeatable ways of working. And when you do things the same way as your rival, you start to look just like your rival—same playbooks, same language, and serving customers through nearly identical interfaces.
Differentiation — the lifeblood of competition — gets cut off by one-size-fits-all static software.
Enterprise leaders are not getting the process efficiency they demand from static software.
The HFS Pulse Study 2025 found process inefficiencies have the single biggest negative impact on the enterprise (see Exhibit 1) when ranked against people, data, and technology challenges. Leaders blame process inefficiencies for limited scalability to support growth or changes in demand (41%), increased operational costs (38%), lower customer satisfaction and retention (34%), slower time to market (31%), and missed opportunities for innovation and process improvement (25%).
Sample: N=305 enterprise leaders
Source: HFS Pulse, 2025
Best practice is a beguiling term. It sounds safe and sensible. But in a volatile world where ambiguity and continuous disruption become normal, ‘best practice’ is yesterday’s assumptions ossified in code: mass-produced logic.
You can’t lead markets with software built for general use cases three years ago. You can’t respond to your customer’s specific context by triggering the same sales actions as your competitor. And you certainly can’t build dynamic, data-driven business models on top of fixed logic and static UI.
SaaS titans Salesforce and ServiceNow recognize this. It’s why Salesforce is so heavily invested in AgentForce – and why ServiceNow is acquiring Moveworks for its agentic AI solution and is also pivoting to an agentic-led approach.
SaaS wasn’t designed for dynamism. It was designed for delivery at scale. It succeeded—but that success came at the cost of uniqueness.
The generative AI revolution has exposed the limits of static software. Now, AI enables real-time orchestration of workflows—not through code hardwired into applications but through intelligent agents and prompts to generate fit-for-context responses and agentic workflow actions.
The software model gets flipped. A third of enterprises (33% in our 2025 Pulse Survey) plan to invest more in autonomous and agentic systems in line with this shift. Half of those supporters are set to increase their investment by 11-20%. One in 10 are ready to increase spend by more than 20%.
Two-thirds of enterprises are frustrated with their SaaS contracts and seek to renegotiate them (see Exhibit 2). Leaders cite feature bloat (new features added that aren’t being utilized) and rigid workflows.
Sample: 605 executives across Global 2000 enterprises
Source: HFS Research Pulse, 2024
Instead of prebuilding logic for every business scenario, AI allows enterprises to generate workflows on the fly, in close-to-real-time, aligned to the specific needs of that moment. This future is a far cry from monolithic software updated once a quarter. It’s a fluid layer—shaped by intent, informed by context, and executed by intelligent agents.
Services as Software (SaS) is a new paradigm in which HFS predicts the majority of services will be delivered by software. To achieve this goal, software, as we know it, must change. Exhibit 3 calls out the major differences between static SaaS and SaS. In SaS, enterprise workflows and interfaces are dynamically generated using AI to interpret intent, understand real-time data, and execute context-aware decisions. Customization is, therefore, infinite, adaptability is real-time, and delivery is autonomous.
You could describe it as ‘workflow as prompt’ or ‘on-demand enterprise logic’, if you prefer. In any event, SaS makes the idea of fixed, static software look increasingly unfit for purpose.
Source: HFS Research, 2025
According to the HFS Pulse 2025 Survey, nearly one in 10 (8%) enterprise leaders report grasping this future—already fully operational or scaling with agentic systems.
In Services-as-Software, the context replaces configuration. Systems no longer require you to define ‘Which predefined process should I run?’ Instead, they ask, “What are you trying to achieve, given the current data, policy, risk, and environment?”
For example, in a supply chain crisis, procurement AI can bypass low-risk approvals and re-route decisions to more agile suppliers based on geopolitical or other filters—without waiting for IT to build a new process. An antidote to twitching tariffs, perhaps?
The enterprise no longer has to adapt to the system; the system adapts to the enterprise.
One example shared with HFS during briefings for the Generative Enterprise Services Horizon 2025 report shows how the responsiveness of agentic workflows delivers improved process efficiency. This case saw the deployment of multi-agent workflows in which LLM agents extract data, evaluate quality, iterate with teams of agents, and qualify with domain experts to deliver compliance–ready regulatory reports to accelerate time-to-market for a global pharma company. The process became more efficient (time cut by three months, at a lower cost), leading to a rise in revenue and earlier access to life-saving drugs for patients.
This shift breaks other things we take for granted. But it also makes room for reinvention.
What breaks:
What gets better:
Services-as-Software challenges conventional governance. Enterprises must move from overseeing static logic to governing intelligent systems that evolve.
So, be ready to:
Provide real-time explainability—to understand why an AI-generated workflow decided what it did
Demand prompt traceability—you will need auditable logs of how actions were generated and completed
Embrace policy-aware models—ensuring compliance constraints are embedded in the model
Offer dynamic entitlements—shifting access rights based on real-time context instead of fixed roles
In SaS, risk, compliance, and trust must be your foundations of orchestration. Without them, no enterprise system will satisfy a regulator.
This future is already emerging in how LLM copilots influence decisions, how orchestration layers rewire workflows, and how data fabrics enable AI to act triggered by real-time signals. It will impact across the enterprise (see Exhibit 4).
SaaS brought predictability to IT budgeting—via annual subscriptions and seat-based licensing. But this model doesn’t map to the AI-native workflows of SaS, which are dynamic, task-based, and often ephemeral.
We’re moving from CapEx-style software—where enterprises buy static capabilities—to Opex-style orchestration, where they consume dynamic intelligence.
The challenge for the CFO is less initial price predictability as costs shift from subscriptions to become tied to tasks and outcomes.
Source: HFS Research, 2025
Leaders across your organization must prepare for change. SaS demands much more than a renegotiated software contract:
Software, as we’ve known it, is dead. Static software optimized predictable processes. We now enter an intelligence-era model of fluid, generative capability to deliver human intent in specific business contexts.
AI changes what software is. The real differentiator won’t be who owns the better software—it’ll be who owns the intelligence to shape what it does, moment by moment.
Embrace this shift from SaaS to SaS, and you will lead the next decade while others keep trying to differentiate with the same software their competitors use—and wonder why they aren’t cutting through.
Register now for immediate access of HFS' research, data and forward looking trends.
Get StartedIf you don't have an account, Register here |
Register now for immediate access of HFS' research, data and forward looking trends.
Get Started