The finance team ran the numbers in January. The SaaS renewal list was shorter than last year — three tools cancelled, two consolidated into one. The procurement team had done its job. So when the annual software invoice arrived in February, everyone expected it to be smaller.

It was not. It was 8% larger. Welcome to the central paradox of SaaS in 2026: the companies cutting the most tools are often the ones watching their bills climb fastest. Understanding why this is happening — and what to do about it — is now one of the most important operational skills a founder, CFO, or IT leader can have.

The Paradox: Fewer Tools, Bigger Bills

The 2026 Zylo SaaS Management Index — the most comprehensive real-world dataset on enterprise SaaS spending, now in its eighth year — tells a story that should be pinned to every IT budget meeting. Enterprise organisations now manage an average of 305 SaaS applications. That number has barely moved: app counts declined just 0.07% year-on-year. Consolidation is happening, but slowly. What is not slow is the spend. Total SaaS expenditure rose 8% year-on-year despite that near-flat portfolio. The average enterprise now spends $55.7 million annually on SaaS — more than last year, with roughly the same number of tools.

For smaller organisations, the numbers are just as uncomfortable. The average company spends approximately $7,900 per employee per year on cloud subscriptions — a 27% increase in two years. Mid-market companies running 100 to 150 SaaS applications are being hit from two directions simultaneously: the cost of the tools they kept is going up, and the tools they are adding to replace them (AI platforms, agentic workflow tools, usage-based services) carry entirely different cost structures that traditional budgeting processes are not designed to handle. Meanwhile, Gartner forecasts global software spending to reach $1.43 trillion in 2026 — growing 14.7%, accelerating from 11.5% in 2025. The market is not shrinking. It is just concentrating into fewer, more expensive, more AI-infused tools.

“SaaS portfolios have flattened, but costs keep rising. Growth is no longer coming from more software — it’s coming from how that software is priced, packaged, and expanded over time.” — Zylo 2026 SaaS Management Index

Why Your Bill Goes Up When You Cut Tools

The conventional assumption about SaaS consolidation is that fewer tools equals lower spend. That was true in the old era of flat per-seat subscriptions. It is no longer reliably true. Three forces are driving spend up even as tool counts decline, and understanding each one is the difference between a consolidation effort that saves money and one that simply reshuffles the bill.

1. AI pricing tiers are being layered onto tools you already own

Every major SaaS vendor — Salesforce, HubSpot, Microsoft, Adobe, Zendesk — has spent the last 18 months embedding AI features into their existing platforms and pricing them as premium add-ons. Salesforce’s Agentforce business hit nearly $800 million in annual recurring revenue in 2026, growing 169% year-on-year. Adobe pulled in $125 million from standalone AI products in a single quarter. HubSpot migrated most of its customer base to a hybrid model that layers AI credits on top of traditional subscriptions. These are not new tools you chose to buy — they are upgrades to tools you already have, often rolled in at renewal with pricing that was not in your original budget. Zylo’s data shows that spending on AI-native SaaS applications jumped 108% year-on-year. For large enterprises, the surge was even more dramatic: 393% in a single year.

2. Usage-based pricing creates invoices nobody predicted

The shift from fixed per-seat licensing to consumption-based models — where you pay per API call, per token processed, per automated action — means that your SaaS bill now fluctuates with usage rather than sitting at a predictable monthly rate. This is better aligned with value in theory. In practice, it means that 78% of IT leaders report unexpected charges from consumption-based or AI pricing models, and 90% of CIOs cite cost forecasting as their top challenge in AI deployment. A tool that looked affordable at procurement becomes expensive at scale. And because usage charges accumulate between billing cycles, the surprise often lands mid-contract rather than at renewal — which is precisely when budget flexibility is lowest. This is directly connected to the broader pricing shift we explored in the move from SaaS to outcome-based pricing models.

3. Consolidation often replaces cheap point tools with expensive platforms

The logic of consolidation is sound: replace five tools with one platform that does everything. The flaw in that logic is that the platform that does everything costs more than one of the five tools it replaced — often significantly more. Mary Meeker’s 2025 Bond Capital report flagged this directly: the era of point solutions is ending, with horizontal platforms like Salesforce dominating. Platform pricing reflects platform ambition. When you consolidate five $50/month tools into one $300/month platform, you have reduced complexity and gained integration — but the bill has not gone down. And given that 61% of organisations were forced to cut other projects or initiatives due to unplanned SaaS cost increases, the downstream impact of this spend inflation is real and measurable.

The Hidden Human Cost of SaaS Sprawl

The financial case for consolidation is well documented. The human case is less discussed but equally compelling. Research from Harvard Business Review found that digital workers toggle between applications nearly 1,200 times per day, spending almost four hours per week — or five full working weeks per year — simply reorienting themselves after switching apps. That is not time spent doing work. It is time spent navigating between the places where work lives.

This context-switching tax compounds in ways that are hard to see in a single line item. Every transition between tools costs cognitive momentum. Every data discrepancy between two tools that should be in sync costs trust and decision quality. Every new hire who needs three weeks to learn the stack costs onboarding time and manager attention. Miro’s own research found that 69% of business leaders say that switching between core work tools and AI tools causes significant friction and interrupts workflows — and that is before accounting for the 37% who struggle to integrate AI with their existing infrastructure at all.

There is also a retention dimension that is easy to underestimate. Talented people do not want to spend their days wrestling with tool friction. When the engineering team has to maintain integrations between eight loosely connected systems to do work that one well-chosen platform could handle natively, the cost shows up in morale, throughput, and eventually attrition. The hidden human cost of sprawl is real. It just does not appear in the SaaS invoice.

What Smart Consolidation Actually Looks Like

There is a wrong way and a right way to consolidate a SaaS stack. The wrong way is to declare “we are cutting apps” from the top down and hand mandates to department heads without understanding how work actually gets done. Consolidation efforts that start with a cost-cutting mandate tend to cut the wrong tools — eliminating tools people actually use in favour of tools that look redundant on a spreadsheet but are load-bearing in practice. The resentment this creates is real, the productivity dip is real, and the shadow IT explosion that follows is predictable.

The right way treats consolidation as a strategic discipline rather than a finance exercise. Leading organisations are asking three questions before making any cuts. First: does this tool create unique, proprietary data or does it just process commodity data that any replacement could handle equally well? Tools that own or generate data that compounds in value over time — your CRM’s historical pipeline, your analytics platform’s long-run behavioural data — are structurally harder to replace than tools that simply transform or display data. Second: does this tool have strong API connectivity and integration depth with the rest of your stack? In an AI-first operating environment, as we covered in our analysis of what AI is replacing inside the SaaS dashboard, the ability for tools to be used by agents — not just humans — is increasingly the dividing line between tools worth keeping and tools worth replacing. Third: can you attribute a measurable business outcome directly to this tool’s existence in your stack?

The enterprises seeing real ROI from consolidation are not necessarily the ones with the biggest budgets or the most aggressive cutting targets. They are the ones who consolidated first and built clean, connected data environments before deploying AI. AI initiatives consistently underperform when built on fragmented tool stacks because AI thrives on connected data and streamlined workflows. When teams are context-switching across eight tools to complete a single process, there is no foundation for AI to augment. This is the strategic dimension of consolidation that most budget conversations completely miss. Consolidation is not preparation for saving money — it is preparation for the AI capabilities that will define competitive advantage over the next three years. This is part of the same structural shift behind the Great SaaS Reset that shook the market earlier this year.

Keep vs Cut: A Framework by Tool Category

What to Do Based on Your Company Size

Consolidation is not one-size-fits-all. The priorities, risks, and starting points differ meaningfully between a 20-person startup, a 500-person mid-market company, and a 10,000-person enterprise. Here is what the data suggests for each.

SMB (under 100 employees) — $14,000 average spend per employee

At this scale, sprawl usually comes from founder-led tool adoption — everyone adds the tool they used at their last company, and nobody retires the ones that overlap. The risk is low governance, high shadow IT, and a SaaS bill that has grown organically without ever being audited as a whole. Start with a full inventory using a tool like Zylo or Torii, identify every subscription and its owner, and ruthlessly cut anything where the owner cannot name a specific business outcome it drives. As a rough guide, an SMB of this size should be running 20–40 core tools maximum. Software spend per employee benchmarks at $14,000 at SMB level — if you are significantly above that without a clear reason, you have optimisation headroom.

Mid-market (100–5,000 employees) — $7,300 average spend per employee

This is where sprawl becomes genuinely expensive and the coordination cost of tool fragmentation starts to show up in business outcomes. Industry data shows mid-sized firms achieved a 29% reduction in applications in 2025 by focusing first on duplicate functionality across departments. The low-hanging fruit alone — three expense tools, five project trackers, four communication platforms — can recover 15–20% of SaaS spend without touching any tool that is genuinely load-bearing. Beyond the cuts, the strategic priority at this size is building a clean data environment ready for AI deployment, which means consolidating around platforms with strong native integration and well-documented APIs. This directly enables the kind of AI agent deployment that is reshaping team structures in 2026.

Enterprise (10,000+ employees) — $55.7M average total spend

At enterprise scale, you are almost certainly managing 400 or more applications, and your challenge is not primarily quantity but governance at scale. The priority shifts from cutting tools to building infrastructure for continuous visibility: who owns each tool, what data does it touch, when does it renew, and is it integrated into SSO and offboarding workflows. Enterprise consolidation that works prioritises platforms with proper APIs and integration capabilities, because every tool that does not connect to others creates expensive middleware requirements and brittle, failure-prone workflows. At this scale, the financial exposure from the SaaSPocalypse market dynamics is also most acute — enterprises are the primary targets of vendor-driven AI tier upsells and consumption pricing changes.

The Shadow AI Problem Hiding in Your Stack

There is a dark irony at the heart of 2026 SaaS consolidation. While IT and procurement teams are focused on rationalising their official stack, a parallel proliferation is happening completely below the governance line. According to research, 67% of employees are now using unapproved AI tools at work. ChatGPT has claimed the top spot in shadow IT charts, displacing the legacy productivity apps that used to dominate that list. 15% of employees routinely use unsanctioned generative AI tools on corporate devices — and 72% of those doing so use personal email accounts to access them, meaning sensitive business data is being processed outside every security boundary your IT team thought it had in place.

This creates a troubling scenario. An organisation can spend six months consolidating its official SaaS stack — eliminating redundancy, tightening governance, cleaning up SSO coverage — and simultaneously be haemorrhaging sensitive data through a dozen AI tools that finance never approved and IT never discovered. The new sprawl is not five project management tools. It is forty-seven employees running Claude, Perplexity, Gemini, and various AI writing tools on their laptops, feeding customer data, internal documents, and proprietary processes into systems that have no data processing agreement with your organisation and no integration with your security stack.

Smart consolidation in 2026 therefore needs a shadow AI component. That means building an AI tool policy — not a blanket ban, which will simply be ignored, but a framework that distinguishes between approved AI tools integrated into SSO, conditionally permitted tools used for non-sensitive tasks, and prohibited tools that touch customer or regulated data. It means treating the AI category in your SaaS management platform with the same rigour you apply to any other software category. And it means recognising that the pace of AI tool adoption among employees will only accelerate, which makes getting ahead of the governance question now significantly easier than cleaning up the mess later.