How AI is transforming SaaS ― and what CIOs should do next

The disruption narrative around AI is not without merit. At its core, AI challenges some of the fundamental assumptions on which enterprise SaaS has been built ­­­­­­­­­­­­­­­­­­― namely, that users should interact with relatively rigid applications, which have predefined workflows and fixed user interfaces to get work done.

AI introduces a different way of working. Users can express intent in natural language, and systems dynamically generate responses, actions and even end-to-end processes without requiring deep knowledge of the underlying application.⁴

This has led to an aspirational vision of a future in which traditional SaaS applications are bypassed altogether. If AI agents can retrieve data, trigger workflows and orchestrate actions across systems, the need for users to log into multiple applications and manually execute tasks appears increasingly redundant. Early use cases from AI assistants generating reports across enterprise resource planning (ERP) and customer relationship management (CRM) systems to agents initiating service requests or financial transactions reinforce this view. This perspective, therefore, assumes that AI can operate independently of the systems that enforce business rules, manage data and enable consistency.

The rapid advancement of large language models (LLMs), combined with enterprise investments in AI-assisted agents, has accelerated the narrative that AI could become the primary interface for enterprise work.⁵

There is also an economic angle to this shift. SaaS vendors have typically monetized through licenses tied to users and functionality. If AI changes how users interact with systems, moving toward consumption or outcomes ­­­­­­­­­­­­­­­­­­― it raises questions about how that value gets priced. At the same time, it’s becoming easier to build AI-driven solutions, which could allow new entrants to compete with established players in more specific areas.

Some organizations have attempted to pivot aggressively toward an AI-led future and have encountered a more complex reality. While early pilots often demonstrate promise, scaling these into enterprise-grade solutions has proven significantly more challenging.

Despite widespread enthusiasm for AI, many organizations continue to struggle with scaling initiatives beyond the pilot stage. Common barriers include workflow integration challenges, limited contextual learning, governance requirements and alignment with day-to-day business operations. While AI adoption continues to accelerate, scaling it responsibly remains a challenge. As organizations move beyond experimentation, gaps in governance, compliance, security and operational oversight continue to hinder the realization of enterprise-scale value.

The EY 2025 Responsible AI Pulse Survey found that nearly three-quarters of organizations have integrated AI across most or all initiatives, yet only one-third have implemented the governance controls needed to manage AI responsibly.⁶ Separately, Work Reimagined research found that while 88% of employees now use AI in their day-to-day work, only 28% of organizations are positioned to convert that activity into meaningful business outcomes.⁷

Beyond these factors, organizations are also finding that AI introduces additional operational and governance complexity. Unlike traditional software, AI models are inherently probabilistic, whereas enterprise environments across industries are designed around deterministic outcomes. For example, credit scoring in a bank is based on defined factors and is expected to produce consistent results: These should not vary depending on how similar information is phrased. In regulated industries such as banking and pharma, this challenge is even more pronounced, as explainability, auditability and compliance are nonnegotiable.

An AI-led interaction layer understands the full context and breaks the request into coordinated actions. Behind the scenes, SaaS platforms act as the control layer ― CRM manages the address change, the claims platform initiates the case and the policy system validates coverage ― supporting maintenance of governance, accuracy and compliance. To enable this, AI accelerates the creation and mapping of APIs that expose SaaS capabilities, allowing workflows to be dynamically orchestrated across systems. At the same time, AI-embedded within each SaaS application enhances the user experience ­­­­­­­­­­­­­­­­­­­― pre-filling data, identifying gaps and recommending next-best actions.

From the customer’s perspective, the experience is unified and conversational receiving a single, guided response that confirms updates, progresses the claim and requests any additional information. Employees are supported with a consolidated, AI-curated view, reducing complexity and improving productivity.

So, this illustrates how these strategic imperatives come together in practice, transforming fragmented processes into a unified, orchestrated model. The next step is understanding how organizations can deliver this at scale.

Each of the strategic imperatives plays a distinct role in enabling this shift right from strengthening core systems to reshaping how users interact with enterprise processes.

Strategic imperative 1: Establish SaaS platforms as the enterprise control layer

To act as an effective control layer, SaaS platforms should be modernized to support API driven execution, resilience and governance at scale. As AI introduces more dynamic ways of working, the need for a stable, governed execution layer becomes even more critical.

Enterprise platforms embed structured processes, enforce business rules and maintain the integrity of critical data ­­­­­­­­­­­­­­­­­­­― capabilities that AI cannot replicate on its own. They also provide the operational backbone that organizations depend on, including audit trails, access controls and enterprise-grade support models with defined service-level agreements (SLAs) and 24/7 coverage. When processes fail or exceptions arise, organizations need more than intelligent recommendations. They need systems that can execute reliably and be supported at scale.

This is particularly important in regulated industries, where outcomes should be consistent, explainable, and auditable. In pharma, for example, platforms such as Pega orchestrate complex clinical and quality workflows, assisting in traceability, version control and compliance with validation requirements.⁸

AI can accelerate individual tasks, be it summarizing clinical data, surfacing insights, or identifying potential safety signals, but it cannot replace the structured workflows, audit controls and deterministic execution required for regulatory submission. AI can enhance decision-making and streamline parts of the process, but responsibility for execution, compliance and accountability remains anchored in these platforms.

As organizations adopt AI, the architectural principle becomes clear: AI can guide and augment, but SaaS platforms enforce and execute. This reinforces the need not just to maintain these systems, but to continue investing in their modernization ­­­­­­­­­­­­­­­­­­­― helping them become resilient, API enabled and capable of supporting AI-driven interaction models.

Leaders should actively modernize and strengthen core SaaS platforms to help them remain a trusted control layer for AI-driven execution.

“The disruption is not that AI makes SaaS irrelevant; it is that AI challenges the assumptions that made SaaS so economically powerful. If agents become the primary interface, the advantage shifts away from application access and toward trusted execution, governed data and outcome-based value,” notes Scott Keipper­­­­­­­­, EY Americas Financial Services Technology Consulting Leader.

Strategic imperative 3: Reimagine user experience through AI‑assisted interactions

Beyond productivity gains, AI is reshaping how users interact with enterprise systems. Traditional SaaS applications have relied on structured, form-driven interfaces and predefined navigation paths. This model is increasingly being supplemented ­­­­­­­­­­­­­­­­­― and in some cases replaced ­­­­­­­­­­­­­­­­­­― by more intuitive, conversational interactions.

Users can now engage through natural language, expressing intent rather than following prescribed steps. Instead of navigating multiple screens, they can ask questions, initiate processes, or request insights more directly, with AI interpreting intent and guiding actions.

This shift does not replace structured processes; it changes how they are experienced. Interfaces become more flexible and intent-driven, while execution remains anchored in governed workflows.

Leading platforms are already demonstrating this shift. ServiceNow’s Now Assist, SAP’s Joule and Pega’s AI-powered decisioning and conversational capabilities enable users to interact with enterprise workflows through natural language. Employees can initiate requests, retrieve insights, or resolve issues without navigating multiple screens, while the underlying platforms continue to execute governed processes.

Over time, enterprise systems will evolve toward more intent-driven experiences, where users focus on outcomes rather than process navigation.

Leaders should redesign user experiences within applications to be more intuitive, AI-assisted and outcome-driven.

Strategic imperative 4: Accelerate integration through AI‑enabled approaches

Integration remains a complex and costly aspect of enterprise SaaS ecosystems. The importance of integration extends beyond technical efficiency. Recent research by the global EY organization ¹¹ suggests that up to 75% of enterprise value remains trapped across organizational and functional silos. Agentic AI can help unlock this value by orchestrating workflows, decisions and actions across systems, enabling organizations to operate around end-to-end value streams rather than discrete applications. Organizations often operate dozens of interconnected systems, requiring significant effort in API development, data mapping and workflow orchestration.

AI introduces meaningful opportunities to reduce this effort; however, it does not eliminate the underlying complexity.

AI-driven tools can automate aspects of integration, including generating API mappings, creating connectors, and supporting low-code orchestration. These capabilities can accelerate implementation timelines, reduce manual effort and lower total cost of ownership. For example, AI can infer data mappings across systems, suggest integration patterns, or generate boilerplate code that developers can refine.

However, core challenges remain ­­­­­­­­­­­­­­­­­­­― data semantics, governance, security and consistency. Integration should still be designed to support traceability, controlled data flows, and clear ownership of system boundaries.

Organizations should treat AI to accelerate integration, helping teams deliver faster while continuing to invest in API-first design, standardized interfaces and modular architectures.

Leaders should adopt AI to accelerate integration while continuing to invest in API-first, well-governed architecture.

Strategic imperative 5: Prepare for AI as the primary interaction layer

A significant shift driven by AI is the gradual disintermediation of the traditional SaaS user interface. As AI agents become more capable, they are increasingly serving as the primary interface through which users interact with enterprise systems.

Instead of logging into multiple applications, users may engage through a unified AI layer that spans systems and processes.

This does not eliminate SaaS platforms. However, it does change how they are experienced. Enterprise applications increasingly operate behind the scenes as API driven engines that execute transactions, enforce rules and maintain records. The user-facing layer becomes more abstract, with AI orchestrating interactions across systems.

This shift should be carefully managed. While a unified AI interface can improve usability and efficiency, it cannot obscure accountability or control. Every action initiated through an AI layer should still be traceable to governed workflows and systems of record.

For CIOs, this shifts the focus from improving individual application interfaces to building platforms that are modular, API-driven and accessible.

Leaders should prepare for AI-led interaction models by making platforms modular, API driven, and ready to operate behind unified AI layers.

The debate around AI and SaaS is often framed in extremes: replacement vs survival. In reality, AI is not eliminating enterprise platforms; it is changing where and how they create value.

Across industries particularly in regulated environments, the fundamentals remain unchanged. Decisions should be explainable, processes auditable and outcomes consistent. When operations affect money, risk, or patient safety, plausibility is not enough.

What is changing is how these systems are experienced and extended. AI introduces more natural, intuitive ways to interact with enterprise processes, accelerating insight and reducing friction. However, it delivers value effectively when embedded within platforms that enforce policy, maintain data integrity, and provide accountability.

“The opportunity for enterprises is in how they reimagine the way they work - their processes, workflows, and decisions ­­­­­­­­­­­­­­­­­­― to streamline experiences for customers and employees, turn phantom AI productivity into measurable efficiency gains, and ultimately make AI a growth lever. That won’t happen through AI and agents alone, but through connecting those agents into platforms and applications that execute for predictable outcomes and predictable costs,” notes Don Schuerman, CTO and Head of Marketing, Pegasystems.

The direction for enterprise leaders is becoming clearer. Organizations should try to focus on the five strategic imperatives ­­­­­­­­­­­­­­­­­­­ reinforcing SaaS systems, embedding AI, evolving user experience, rethinking integration and preparing for new interaction models.

The approach overcomes the “inertia moat” identified earlier that organizations across sectors are facing. It enables organizations to build on existing platforms that already meet regulatory and operational needs, rather than replace them amid uncertainty around AI cost and risk. As a result, adoption is likely to progress incrementally with AI being embedded into existing systems to enhance, rather than disrupt, enterprise workflows.

In many ways, this moment mirrors previous technology shifts. Just as cloud did not eliminate enterprise software but reshaped how it was delivered and consumed, AI is redefining how SaaS platforms are built and experienced. The destination is not fewer platforms but more capable ones ­­­­­­­­­­­­­­­­­­­― AI-enabled, API driven and designed to operate within broader orchestration layers.

The invention of autopilot did not eliminate pilots ­­­­­­­­­­­­­­­­­­­― it changed their role. The question is no longer whether AI will reshape SaaS, but how quickly organizations can respond. Those who move decisively ­­­­­­­­­­­­­­­­­­­― strengthening platforms while embedding AI in a controlled, scalable way ­­­­­­­­­­­­­­­­― will define the next generation of enterprise software.

Summary

AI is reshaping how enterprises engage with SaaS, but these core platforms remain central to governance, compliance and execution. The shift is incremental, moving from application-centric workflows to more intuitive, AI-driven experiences, with SaaS platforms operating as the backbone for decision-making and control. Organizations that focus on strengthening their existing systems, embedding AI capabilities and preparing for evolving interaction models will be better positioned to unlock value. The evolution ahead is not about eliminating platforms, but making them more intelligent, connected and adaptable.