← Back to blog
SaaS & Integraties

The average organisation now runs 830 apps: why AI tools are loosening your grip on software, not tightening it

By Zarioh Digital Solutions5 min read
Share

A new international SaaS benchmark study shows that organisations now run more than 830 applications on average, and that over six in ten fall outside IT's view entirely. AI tools are the fastest-growing driver. What does that mean for your overview, your data, and your integrations?

An employee opens an AI writing assistant, signs in with their work Google account, grants it access to documents and mail, and is working within a minute. Nobody bought anything, nobody approved anything, and IT has no idea it exists. This scenario plays out dozens of times a day inside every mid-sized organisation, and it is exactly why the corporate software landscape is growing faster than ever.

An international benchmark study covering thousands of organisations, published earlier this year by SaaS management platform Torii, lays out the numbers. The average organisation now runs more than 830 applications, with a median of 680. At large enterprises that figure climbs past two thousand. And of all those apps, 61 percent sit outside formal IT oversight. Only about 15 percent are actually approved and deliberately purchased.

AI was supposed to consolidate. It is doing the opposite

A few years ago the expectation was that AI would help consolidate software: fewer tools needed, because one assistant could do the work of five separate apps. Reality has gone the other way. The research shows AI-first tools are now the dominant growth engine behind new, unmanaged software adoption. More than half of the most widely used shadow applications inside organisations today are AI tools: meeting note-takers that listen in, writing assistants, code generators, email and document summarisers.

The reason is simple. Where a traditional SaaS tool used to require a credit card, a contract, or a conversation with procurement, an AI tool usually needs just one click on 'sign in with Google' or 'sign in with Microsoft'. No invoice, no approval, no conversation. The friction that used to force a question, do we already have this, and should this tool really see our data, has disappeared.

The real risk is not the number of apps

Hundreds of apps are not a disaster on their own. The problem is that nobody knows exactly which ones, who has access, and what they can see. Any AI tool connected through a simple OAuth grant to a work account can potentially reach mail content, calendars, shared documents, and contact lists. That access frequently stays active long after an employee has stopped using the tool, or even after that person has left the organisation. An account nobody actively uses anymore but that still holds permissions is exactly the kind of gap that only surfaces after an incident.

On top of that, these tools often connect to each other. An AI note-taker forwards summaries to a separate project management tool, which syncs with a CRM running on a free account set up by someone who has since changed roles. At every link in that chain, it is unclear who is actually responsible for the data flowing through it.

Blocking everything backfires

The instinctive response, blocking every new tool until IT approves it, rarely solves much in practice. Employees who find an AI tool genuinely useful for their work will find a way around the block: a personal phone, a private account, a browser outside company policy. The problem does not disappear, it just disappears from IT's view, which is exactly where the shadow IT numbers come from. A hard block feels safe but in practice often grows the part of the software landscape nobody is watching.

A workable approach starts with visibility, not a ban. Most organisations can pull a list of every app currently connected through a signed-in work account from the Microsoft 365 or Google Workspace admin centre within an hour. For most companies, that is the first time anyone actually sees how many connections are genuinely active, separate from whatever was formally purchased at some point.

From scattered connections to one overview

The next step is structural: centralising integrations instead of letting every tool set up its own connection. A central integration platform, where data moves between systems through one managed layer instead of dozens of individual OAuth grants, restores visibility over what comes in and where it goes. That is exactly where many organisations are stuck today: the tools themselves are fine, but the way they connect to each other and to company data was never deliberately designed.

A light, fast approval route for new software also helps more than a heavy one. If an employee gets an answer within a day on whether a tool can be used, including a quick check on what data it can see, there is no reason left to work around that process. The goal is not to block every tool, the goal is to know what is actually running.

Three steps to start with tomorrow

First, pull the list of connected apps from your Microsoft 365 or Google Workspace admin centre and review it with IT and a few team leads. Many of the surprises are not the big, well-known tools but small AI helpers someone tried three months ago and never stopped using. Next, set a simple rule: new software requesting access to company data gets reviewed within a working day, not a month. Finally, bring your most important connections, especially between core systems such as accounting, CRM, and communication, into one managed integration layer instead of loose connections nobody can fully account for anymore.

Getting a grip on your software landscape does not mean using less AI. It means knowing what is coming in, who can reach it, and how your systems actually talk to each other. Want to know which apps and connections are genuinely active inside your organisation, and how to structure that properly? Zarioh helps organisations with SaaS management and integrations that are deliberately designed, not accidental.

Z

Zarioh Digital Solutions

IT specialists from Utrecht, the Netherlands. We help businesses with Microsoft 365, AI agents, hosting and telephony, and share what we learn in practice. Follow us on LinkedIn

Related articles

← Back to all articles
Share