Walk into any modern office — physical or virtual — and you’ll find layer upon layer of software quietly humming beneath every business decision, customer interaction, and internal process. From the CRM logging a sales call to the payroll tool processing Friday’s wages, software applications have become the operating system of the modern business. But just how many applications does the average company actually rely on? The answer is larger than most people expect — and the landscape is shifting faster than ever.
The Numbers: A Brief History of Accumulation
The rise of cloud-based, subscription software — known as Software as a Service, or SaaS — transformed how businesses buy and use technology. Instead of purchasing expensive, rigid systems, companies could simply sign up for a tool, pay monthly, and add more as needs emerged. The result was an explosion of applications.
Between 2015 and 2022, the average number of SaaS applications used by businesses grew by a staggering 1,525%. This wasn’t a fringe trend — it reshaped entire IT departments and spawned new professions dedicated purely to managing the software stack.
By company size, the differences are pronounced. Smaller companies with under 200 employees use an average of 42 SaaS applications, while large organizations with over 5,000 employees report an average of 158. At the very top end of the scale, large organizations with more than 10,000 employees use around 447 SaaS apps on average.
The Peak and the Pullback
After years of uninterrupted growth, something changed. After peaking in 2022 at 130 SaaS applications on average per company, the latest data reveals that the number dropped by 14% to 112 in 2023 — the first decline in over a decade. Since the 2022 peak, usage has fallen 18%, marking the second consecutive year of decline as businesses cut back on non-essential tools.The reason isn’t disillusionment with software. It’s discipline born from economic pressure. Over half of respondents in recent surveys felt there was more scrutiny in SaaS purchasing than before, with companies reporting wasting on average more than $135,000 in unused software licenses. Excess had become expensive. Studies show that 53% of SaaS licenses go unused within 30 days — driving major waste.
This consolidation phase reflects a maturing market. Businesses aren’t abandoning software; they’re becoming more deliberate about which software earns its place. The SaaS market is growing fast, but companies are using fewer apps. They’re cutting out weak tools and sticking to fewer, stronger platforms that do more.
The Hidden Problem: Shadow IT
Even the official count of applications understates reality. Many employees adopt tools independently — without IT’s knowledge or blessing. As of 2025, 48% of enterprise apps are shadow IT apps, meaning software employees use without the IT department’s knowledge or approval.This creates compounding risks. 56% of employees upload sensitive information into applications that are not approved, and IBM’s 2024 report found that one in every three data breaches now happens because of shadow IT. The sheer volume of applications — official and unofficial — has made governance a significant challenge in its own right.
What Businesses Are Actually Running
The applications businesses use span virtually every function. Customer relationship management (CRM), HR and payroll, project management, communication and collaboration, accounting and finance, marketing automation, data analytics, cybersecurity, and document management all form the core of a typical enterprise stack. Tools like Slack, Microsoft 365, Salesforce, HubSpot, Zoom, and Google Workspace are near-ubiquitous. Based on 2024 expense data, the most frequently renewed SaaS tools include ChatGPT, Canva, LinkedIn, Udemy, Grammarly, and Adobe Acrobat.
The typical individual department doesn’t escape this complexity. The average department in an organization uses about 87 SaaS applications — a figure that surprises most people outside of IT.
How the Landscape Will Change
The next chapter for business software is being written by artificial intelligence — specifically, autonomous AI agents capable of reasoning, deciding, and acting without constant human oversight.Unlike traditional SaaS applications that require users to click through interfaces, fill out forms, and manually execute workflows, AI agents operate as autonomous systems capable of reasoning through problems, making decisions, and taking action without constant human oversight. They understand natural language commands like “analyze our Q2 performance” — eliminating the need for users to learn complex navigation paths through multiple applications.This shift has major implications. Instead of navigating multiple dashboards, users could interact with agent-driven, conversational interfaces that perform tasks across systems — instructing an AI agent to “approve last week’s expense reports” or “generate next quarter’s sales forecast” and having the agent orchestrate workflows across HR, finance, and CRM systems behind the scenes. In this model, the number of applications a person consciously interacts with could shrink dramatically, even as the underlying infrastructure remains complex.
In three years, any routine, rules-based digital task could move from “human plus app” to “AI agent plus API.” Traditional SaaS vendors are acutely aware of this pressure and racing to embed AI capabilities into their platforms before upstart, AI-native competitors displace them.However, analysts caution against overestimating the speed of disruption. Deloitte predicts that the full replacement of enterprise applications by agents won’t happen in 2026 — it will likely take at least five years or more to come to fruition, even with the rapid pace of technological development. Traditional SaaS providers have large footprints across complex workflows that will be hard to supplant.
Pricing models will also undergo a fundamental shift. IDC predicts that by 2028, pure seat-based pricing will be obsolete, with 70% of software vendors refactoring their pricing strategies around new value metrics such as consumption, outcomes, or organizational capability. In other words, businesses may stop paying for software per employee and start paying for software per outcome.”AI isn’t going to trigger a ‘SaaSpocalypse’ so much as a ‘SaaSmorphosis,'” according to future of work economist Richard Johnson. “They both can coexist. However, the ‘S’ in SaaS that changes isn’t the software but the service.”
The average business today juggles over 100 software applications to keep its operations running — and large enterprises manage many times that. After a decade of accumulation, a period of consolidation is underway, driven by budget discipline and a desire for integration over proliferation. But the more profound transformation lies ahead: AI is poised to reshape not just how many applications businesses use, but what software fundamentally *is* — shifting it from a collection of tools employees navigate to an intelligent layer that works on their behalf.
For business leaders, the question is no longer just “which software do we need?” but increasingly “what should software actually do for us?” Those who answer that question well will be the ones writing the next chapter of this story.