5 Industry Experts Weigh In on the 2026 SaaS Management Index
Table of Contents ToggleTop 15 Most Expensed Apps for 2025What Can...
Back
Back
Search for Keywords...
Blog
Table of Contents
Updated on February 20, 2026 with new stats and further analysis on shadow IT and AI.
As SaaS evolves, integrating with AI, employee enthusiasm for choosing their own tools remains high. In fact, Zylo’s 2026 SaaS Management Index shows that 3.4% of employees expense SaaS. Many of the applications they’re expensing contain AI features or are AI-native apps—those built on AI, like ChatGPT, Anthropic, and Perplexity.
Expensed apps—or shadow IT—are a source of uncontrolled spend, SaaS sprawl, and security risks. Our data shows that 59% of expensed apps have “Poor” or “Low” security ratings via Netskope’s Cloud Confidence Index™. Expensed apps also result in nearly two million dollars in costs per year on average per company. This spending makes up only 3.7% of total SaaS spending but still accounts for 45% of applications.

These applications offer convenience, flexibility, and cost-effectiveness, making them an attractive choice for employees. However, the risks of shadow IT should remain a concern for IT leaders.
According to Zylo’s 2026 SaaS Management Index, the most expensed apps from 2025—based on transaction volume—were:
Applications like Kudoboard, CliftonStrengths, and Coursera did not make the list this year, signaling a shift in employee software preferences. The biggest driver of that change is AI.
In 2023, we saw AI-native apps appear on the list of most expensed apps for the first time. Now, ChatGPT takes the #1 spot, up from #2 last year, and a giant leap from #14 in 2023. Beyond ChatGPT, one noticeable difference this year is the growing prevalence of expensed AI tools. Zylo data also showed that eight out of the top 50 most expensed applications are AI-native apps—or 16%.
What we’re seeing is the introduction of a new source of hidden spend and risk: shadow AI. It’s like shadow IT in that AI applications are being purchased outside of sanctioned channels, without IT’s awareness.
As if managing hidden costs and risks of shadow IT weren’t hard enough, AI adds new complexities. How do you keep tabs on consumption and budget appropriately? How is your data used and stored by the vendor?
Whether a SaaS app with AI features or an AI-native app, you need to bring unknown expense purchases to light. While apps differ from company to company, let’s examine the biggest culprits.
Among the top five most expensed SaaS applications are three newcomers: Apple iCloud, QuickBooks, and OpenAI API. Canva remained in the #3 spot, while ChatGPT moved up just one place and OpenAI API two. Meanwhile, Apple iCloud made the largest leap from #13 to #2. QuickBooks is a first-time addition to the most expensed apps list.
Here’s a breakdown of the five most expensed apps in 2025.
ChatGPT is a generative AI tool where users interact with an AI bot or agent, entering prompts to create images, text, and/or video via AI. Use cases range from conducting research and content creation to data analysis and coding. As employees seek to become more efficient in their work, we’ll continue to see them expense AI apps. Governance and education will be critical to ensure data protection and proper use of AI tools.
iCloud is Apple’s cloud storage and synchronization service, enabling users to store files, back up devices, and sync photos, email, and documents across Macs, iPhones, and iPads. In the workplace, employees often use it for file sharing, large media transfers, or backing up work created on Apple devices. Because upgrades to paid storage tiers are inexpensive and tied to personal Apple IDs, employees frequently expense iCloud plans outside standard procurement workflows.
Canva is a graphic design platform allowing users to create professional-looking designs for various purposes, such as social media posts, presentations, and marketing materials. The platform offers a range of images, templates, and design elements, making it easy for users to create visually appealing content without prior design experience. Democratizing design, plus the addition of AI features the last few years, has made it easier to scale creative work—especially for those without a design background.
QuickBooks is accounting software designed to help businesses manage invoicing, payroll, expense tracking, and financial reporting. Teams often adopt it for departmental budgeting, contractor payments, or managing side business operations tied to company initiatives. Because it’s easy to purchase with a credit card and set up without IT involvement, employees may expense QuickBooks subscriptions directly—especially in distributed teams where finance workflows develop outside centralized procurement processes.
OpenAI API allows developers to integrate generative AI capabilities directly into applications and internal workflows. Product, engineering, and data teams often use it to prototype AI features, automate tasks, or power customer-facing experiences. Because usage-based billing is tied to API keys and company credit cards, teams may expense OpenAI API costs directly, especially during experimentation phases outside formal procurement channels.
In 2025, spending on artificial intelligence skyrocketed. Zylo data found that AI-native application spend rose 108% on average—and nearly 400% for large enterprises. Apps with the highest spend growth included:
However, as more SaaS companies add AI features and functionality, the line gets blurred between what is AI and what’s not. According to the 2025 SaaS Benchmarks Report by High Alpha, 64% of SaaS companies say AI is now a supporting feature in their tools. Meanwhile, 36% say AI is core to the product itself. The line is actually clearer than you think: nearly all software now contains or is built on AI.
While helpful for productivity and innovation, the unauthorized use of AI apps and features can lead to hidden costs, such as subscription fees or API usage charges. Also, many are monetized through usage-based or hybrid models, which make costs volatile and hard to predict.
There’s also significant data security concerns with using GenAI tools like ChatGPT, as employees may inadvertently share sensitive information with the tool. Of the IT leaders we surveyed, 43% said their biggest concern around the use of AI is exposure of sensitive company data. Regulatory and compliance risks around AI use (33%) came in second.
For more data like these, check out our list of 175+ unmissable SaaS stats, or download our 2026 SaaS Management Index.
Effective SaaS Management requires continual discovery, active monitoring, and identification of all new SaaS expense purchases. To address expensed SaaS:
Effective SaaS Management requires continual discovery, active monitoring, and identification of all new SaaS expense purchases. To address expensed software:
You can’t manage what you can’t see—and that includes shadow IT. Companies frequently underestimate the number of SaaS in their environment. For instance, the former Sr. Software Asset Manager at Genesys, Samantha Griffin, shared that they initially thought they had 100 apps in their environment. After going through the discovery process, they uncovered 600—500% more.
Without knowledge of hidden SaaS applications, you’re likely introducing unnecessary risk and spend. Having a clear understanding of your SaaS stack can inform what additional security measures and spend controls you may want to put in place.
For example, Zylo’s AI-powered financial discovery continuously identifies applications from expense and AP data, helping enterprises establish a comprehensive SaaS system of record.
How the Zylo Discovery Engine Powers the Most Comprehensive SaaS Management Platform
Learn MoreThe Freedom within a Framework SaaS governance approach makes it easy for employees to select their preferred tools within specified guardrails. Adopting this approach helps:
Start by creating policies to curb rogue software purchases. Often, that involves putting restrictions on expense purchases—either setting a dollar threshold for what may be expensed or banning them altogether. The route you choose depends on what makes sense for your business.

By addressing the underlying causes of shadow IT, businesses can ensure better control over their SaaS portfolio and maintain a secure and efficient IT environment.
Evolving Your SaaS Governance Framework for the Digital Workplace
Learn MoreFor many organizations, introducing a formal software review and approval process can keep expense spending in check. It can be as simple as an application request form to a comprehensive software review board.
Determine what information is required for a software request, and build it with an existing tool, like Google Forms or Typeform. Then, the key is to make the form easily accessible across the business. For instance, add it to your company intranet or another location employees frequently visit.
Form a cross-collaborative team, including members from Procurement, IT, Legal, Security and Accounting. Determine the process you’ll follow to review new application requests, clearly identifying each stakeholder’s role.
The potential downside is that a review board may slow down the acquisition process and restrict who can acquire software. However, many Zylo clients say it effectively reduces shadow IT and expense spending.
Once you have governance in place, you can begin to eliminate shadow IT by following these best practices:
Start by identifying which apps you already have an enterprise agreement for. Then, check if licenses are available and migrate the user or users under that contract. Before canceling the other subscription, understand the terms and when you’re able to pull the plug.
If you don’t already have an enterprise agreement, notify end users they’ve made a purchase outside of policy and request they cancel the subscriptions. Next, direct them to your application catalog to select an already approved tool. If they’re unable to find a tool they need, provide direction on how to request the purchase of a new tool if that fits their needs better.
At this point, you’ve done all this hard work, the last thing you want is for shadow IT to creep back into your environment. Make sure you have proper monitoring in place so you’re alerted when a new app or payment comes in that is against your policies. Then you can take quick action, whether asking to cancel the subscription or not reimbursing them on future payments.
The rise of shadow IT and shadow AI highlight the need for organizations to gain a comprehensive understanding of their SaaS landscape. With complete visibility, effective governance, and decisive action, your organization will reduce expense spend and compliance risks.
Do you know where your organization’s biggest opportunities to cut down on shadow IT are? By pairing Zylo’s AI-powered financial discovery with our SaaS Management methodology, you can save money and improve fiscal accountability across your organization. Schedule a demo and see how Zylo can go to work for you.
An expensed SaaS or AI application is software purchased directly by an employee using a corporate card or personal reimbursement, rather than through centralized IT or procurement processes. These tools often bypass formal review, creating visibility gaps in cost, ownership, security, and contract management.
Employees often purchase SaaS and AI tools independently because they are easy to access, inexpensive to start, and deliver immediate value. Lengthy approval processes or unclear purchasing policies can also encourage teams to expense tools directly to move faster.
Shadow IT refers to expensed SaaS applications purchased outside formal procurement channels. Shadow AI specifically involves expensed AI tools, such as generative AI platforms or APIs. Both create visibility and governance gaps, but shadow AI often introduces additional data security and usage-based cost risks.
Shadow IT and shadow AI can expose sensitive data, bypass identity and access controls, and introduce applications that haven’t been vetted for compliance. AI tools may also process or store proprietary information in ways that increase regulatory, contractual, or intellectual property risk.
IT teams can uncover expensed SaaS and AI tools by analyzing accounts payable and expense data, monitoring corporate card transactions, and using automated discovery. For example, Zylo’s AI-powered financial discovery continuously identifies applications from expense and AP data, helping organizations build a complete, centralized SaaS system of record.
Yes. Although expensed SaaS represents only 3.7% of total SaaS spend, it accounts for 45% of applications. That volume increases oversight costs, fragments vendor negotiations, and drives redundant renewals—ultimately inflating IT operating expenses and reducing budget predictability.
Expensed applications often create duplicate contracts, inconsistent pricing, and unmanaged license counts. Without centralized oversight, renewals may auto-renew at higher prices, and unused licenses can go unnoticed, complicating optimization efforts and reducing overall purchasing efficiency.
Procurement can enforce stronger SaaS policies by establishing clear approval workflows, setting spend thresholds, and educating employees on approved tools. Providing pre-vetted alternatives and centralizing contract ownership helps reduce off-channel purchases while maintaining business agility.
ABOUT THE AUTHOR
Nicole Wood
Nicole Wood is the Senior Content Strategist at Zylo, where she develops content that educates and empowers enterprises to manage SaaS strategically. She is also the producer the Silver Stevie Award-winning podcast, SaaSMe Unfiltered.
Table of Contents ToggleTop 15 Most Expensed Apps for 2025What Can...
Table of Contents ToggleWhat Is SaaS Compliance?Why SaaS Compliance MattersKey Types...
Table of Contents ToggleTop 15 Most Expensed Apps for 2025What Can...
Table of Contents ToggleTop 15 Most Expensed Apps for 2025What Can...
| Cookie | Duration | Description |
|---|---|---|
| cookielawinfo-checkbox-analytics | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Analytics". |
| cookielawinfo-checkbox-functional | 11 months | The cookie is set by GDPR cookie consent to record the user consent for the cookies in the category "Functional". |
| cookielawinfo-checkbox-necessary | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookies is used to store the user consent for the cookies in the category "Necessary". |
| cookielawinfo-checkbox-others | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Other. |
| cookielawinfo-checkbox-performance | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Performance". |
| viewed_cookie_policy | 11 months | The cookie is set by the GDPR Cookie Consent plugin and is used to store whether or not user has consented to the use of cookies. It does not store any personal data. |