Vijay Raina is a preeminent voice in the world of enterprise technology, specializing in the evolution of SaaS and its underlying architecture. As global organizations grapple with the sudden rise of autonomous intelligence, Vijay provides the strategic framework needed to navigate the transition from traditional software applications to agentic workflows. His insights help CIOs and CTOs understand that the current shift isn’t just a technical upgrade, but a fundamental redefining of how value is created and measured within a digital business.
The following discussion explores the profound pressure points currently hitting the SaaS model, specifically regarding how pricing is shifting from user seats to completed outcomes. We delve into the changing nature of user interfaces, which are transforming from active workspaces into centralized command centers, and the strategic battle for the orchestration layer that sits above individual applications. Vijay also outlines the necessary steps for data governance and workforce retraining to ensure companies are prepared for a future where agents, not just humans, are the primary operators of enterprise software.
The shift from per-user licenses to outcome-based pricing is a radical departure from how we’ve bought software for decades. How do you see this fundamental change affecting the financial stability and budgeting of global enterprises over the coming years?
The shift from per-seat metrics to outcome-based pricing represents a seismic relocation of the unit of value in enterprise technology. For decades, the industry relied on the assumption that humans were the primary unit of work, but agents can now break that assumption by performing complex tasks like invoice reconciliation overnight without human intervention. We are seeing a move toward hybrid models that combine traditional licenses with value-based components to avoid the “bill shock” of unpredictable token or inference costs that often accompany scaled AI. According to recent data, 74% of companies expect to integrate agentic AI moderately within the next two years, forcing a radical rethink of how software budgets are allocated. This isn’t just a cost adjustment; it’s a strategic move where vendors stop selling software “seats” and start selling “reconciled invoices” or “resolved tickets.”
As AI agents take over the heavy lifting of reasoning through multi-step tasks, you’ve suggested that the traditional user interface is becoming more of a “control tower.” What does this mean for the daily experience of the average enterprise employee?
As agents take over the heavy lifting of reasoning through multi-step tasks, the classic dashboard is evolving into a centralized digital command center where humans act as supervisors rather than operators. We are moving toward a headless architecture where the direct interaction with a specific software platform becomes secondary to the orchestration layer that sits above it. Instead of spending hours navigating menus and inputting data across various portals, employees will spend their time setting high-level goals and intervening only when specific human judgment or exception handling is flagged. This transition demands that we view the application not as a destination, but as a utility that reports into a broader, neutral orchestration layer. It essentially puts people back onto the highest-value problems by removing the friction accumulated from decades of layered SaaS investments.
Large enterprises are unlikely to rely on a single vendor for all their AI agents. How should a CTO decide whether to bet on a single vendor’s ecosystem or invest in building a neutral orchestration layer?
This is perhaps the most critical architectural decision a CTO will face in the next three to five years because large enterprises will likely never run on a single vendor’s suite of agents. You will have agents from incumbent providers, AI-native startups, and those built in-house, all needing to work together seamlessly across finance, customer service, and IT operations. Choosing to lock into one vendor’s walled garden offers ease of integration today but risks losing the transparency and portability of data that is essential for long-term governance. Organizations should be deliberately considering a federated model or a neutral orchestration layer to ensure they do not find themselves having to live with restrictive architectural consequences later. The focus must remain on who owns the “control plane” that manages these agents, as that is where the true power and efficiency of the modern enterprise will eventually reside.
You’ve noted that many organizations overestimate their data readiness for this new era. What are the specific prerequisites for a data foundation that can actually support autonomous agents acting on behalf of the company?
The gap between perceived readiness and reality becomes painfully clear the moment an agent begins acting on data that lacks proper lineage, observability, or governance. To deploy agents safely at scale, companies must establish a foundation that prioritizes access and visibility, ensuring that every action taken by an agent is traceable and authorized within the company’s existing security frameworks. Beyond the technical side, it is crucial to renegotiate the very nature of vendor relationships by building “agentic clauses” into contract renewals that protect data portability and the right to integrate with non-vendor orchestration layers. Without these structural safeguards, the agents will struggle to navigate the “sticky data” environments of traditional SaaS providers, leading to errors and operational silos. Truly leading enterprises are treating this as a structural shift to architect for, rather than just another minor feature to consume from their existing tech stack.
This transition is often described as a cultural shift rather than just a software upgrade. How do we effectively retrain a workforce that has been trained for years to “use” software to now “supervise” a team of agents?
We have to stop thinking about our employees as software users and start training them as agent supervisors who manage a hybrid workforce of humans and machines. This requires a fundamental pivot in skill sets toward validating outputs, setting strategic goals, and knowing precisely when to step in with human intuition during a multi-step task. It is a massive cultural undertaking because it changes the day-to-day rhythm of work from manual processing and data entry to high-level oversight and exception management. As 74% of the enterprise landscape moves toward this model, the companies that succeed will be those that foster a culture of supervision and critical thinking rather than simple execution. The goal is to eliminate the friction of legacy workflows, allowing the workforce to focus on solving the complex, non-algorithmic problems that define a competitive edge.
What is your forecast for the SaaS industry over the next five years?
In the next five years, I expect the “SaaS as an application” category to largely dissolve, replaced by a landscape where the primary value is the autonomous action rather than the software seat. We will see a consolidation of the orchestration layer where a few dominant “command centers” manage a vast ecosystem of specialized agents across every corner of the enterprise, from the supply chain to HR. Organizations that fail to build a robust data foundation today will find themselves hampered by unpredictable costs and rigid architectures that cannot adapt to the speed of agentic work. Ultimately, the successful enterprise will be one that has fully integrated a hybrid human-agent workforce, where software is no longer a tool to be operated, but a partner that executes the heavy lifting under human guidance. This window to redesign work is open now, but it will not stay open for long as the momentum of Agentic AI continues to accelerate.
