Across industries, companies are transitioning from AI pilots to real-world deployments and achieving measurable results. Teams are generating insights, completing research, and drafting work in a fraction of the time. What once took months now takes days—what once took days now takes minutes.
But then, the momentum slows.
That’s when companies realize that productivity gains alone don’t automatically lead to business transformation.
McKinsey found that 62% of organizations are experimenting with AI agents, but only 39% have reported a financial impact. Gartner predicts that by the end of 2027, more than 40% of agentic AI projects will be canceled.
The reason is simple: a key part of the workflow hasn’t changed. Organizations still rely on people to bridge the gaps between systems.
People as middleware
For decades, enterprise workflows have been designed around people, with humans manually transferring information between different applications. Each type of software—CRM, ERP, project management platforms, productivity suites—was built for one kind of user: someone who must manually connect processes, because these applications themselves cannot fully integrate.
People switch back and forth between these systems, copying information from emails, and clicking through countless interfaces and other tools. I call this the “toggle tax.” Researchers at Harvard Business Review found that employees switch between applications about 1,200 times per day, with merely re-entering states and context switching consuming nearly 9% of the workweek.
Rethinking the Role of SaaS
This is a heavy and time-consuming cost, and agents are uniquely positioned to eliminate it. This prospect has sparked predictions of a “SaaSpocalypse.”
This point is straightforward: if agents can access the same data, the obvious question is—why do we still need this application?
I think a better question is, what will the application become?
In an agent-driven world, SaaS will no longer be something people log into, browse, and switch between. It will evolve into a system that provides agents with the unique context they need to act reliably and securely. Its value has never been the interface—it has always been the underlying data, relationships, governance, business rules, and logic.
As agents coordinate work across applications, people can spend less time moving information and more time serving customers, making decisions, and driving business outcomes. Gartner predicts that by 2028, 33% of enterprise software applications will include agentic AI, compared to less than 1% in 2024.
Redesigning work for the agent era
Achieving future opportunities requires more than just deploying new technologies. Microsoft’s latest Work Trend Index shows that only one in four employees believe their leadership is aligned on AI. The productivity J-curve explains why this matters: transformative technologies can only unlock their maximum benefits after organizations align around a shared vision and redesign the workflows that support them.
Leaders don’t need to overhaul their business overnight to move forward; in many cases, the next step is simpler than they think. Start with a workflow that spans multiple applications and delivers clear business value. Improve it, learn from it, and then scale up. As momentum builds, identify other workflows still slowed down by toggle tax and simplify them in the same way.
The more powerful the agents, the more reliant they become on interconnected enterprise applications. But their greatest impact comes from organizations redesigning processes around this new way of working. As this evolution takes hold, people will step away from the role of "human glue"—no longer responsible for coordinating between systems, but instead focusing on what only humans can do: making decisions, building relationships, and driving business forward.
The views expressed in the Fortune.com article are solely those of the author and do not necessarily reflect the views or positions of Fortune.