Shadow AI Is the New Shadow IT - What Calero's Consumption-Cost Release Means for Your SaaS Spend

Shadow AI Is the New Shadow IT: What Calero's Consumption-Cost Release Means for Your SaaS Spend

Most IT and finance leaders already know the feeling of opening an invoice and finding a subscription that nobody remembers approving, and that problem is about to get harder because artificial intelligence is entering the enterprise through free tiers, individual signups, and features quietly switched on inside software you already license. On September 16, 2026, Calero, the expense-management platform behind the Cortexa Experience (Powered by Calero), announced a set of SaaS Management capabilities built specifically to find that spend and control it before it compounds. We think the release matters less as a product story and more as a signal about where technology spend is heading, so here is what changed and what it means for your budget.

What did Calero actually announce?

Calero, which describes itself as a Leader in the 2026 Gartner Magic Quadrant for SaaS Management Platforms, extended its platform to manage AI consumption, cost, and adoption alongside the telecom, mobility, and SaaS spend it already tracks in a single view. The release adds five capabilities that work together across the same platform. Shadow AI Discovery scans the systems already running in your environment and classifies every AI or AI-enabled tool by vendor and function; Hybrid License Management pulls seat-based licenses and consumption-based entitlements, including pooled tokens and individual quotas, into one picture; Adoption Tracking flags idle seats and underused consumption plans; Anomaly Detection lets teams set token thresholds and fire alerts on usage spikes before the monthly bill lands; and pre-built integrations collect provider-specific data from Microsoft Copilot, Cursor, and Anthropic Claude in one place. Eric Martorano, Calero's President and CRO, framed the shift bluntly, noting that AI is arriving with a meter running.

Why is shadow AI harder to catch than shadow IT?

Shadow IT was fundamentally a seat-count problem, and a seat count is at least a stable number you can reconcile against a contract at renewal time. Consumption-based AI breaks that math, because a tool priced on tokens or usage can swing from a rounding error to a five-figure line inside a single billing cycle, and the first warning most teams receive is the invoice itself. The other complication is the entry point, since AI rarely arrives as a signed enterprise agreement and instead shows up as a free trial a manager started, a Copilot feature toggled on inside a license you already pay for, or a developer expensing a coding assistant on a corporate card. When half the entitlement picture is missing, renewal and true-up decisions fail in predictable ways, which is exactly the gap that continuous discovery and hybrid-license tracking are built to close.

How much SaaS and AI spend is actually hidden?

The scale of the problem is now well documented across independent research. BetterCloud's 2026 State of SaaS report found that the average mid-market organization watched its application count jump 41 percent in a single year, climbing from 116 apps to 164, and that only 56 percent of the applications in active use carried formal IT approval. In our own engagements the visibility phase routinely surfaces 30 to 40 percent more SaaS applications than a client believed they were running, and much of that spend renews on autopilot against a budget nobody owns. Layer variable AI consumption on top of that baseline and the uncounted portion of the technology estate grows faster than any annual review cycle can realistically track.

What does consumption pricing break in your renewal strategy?

For a finance leader the risk is no longer a predictable line you can forecast and negotiate once a year; it is a variable that compounds quietly in the months between reviews. We have written before about how Zoom's move to a paid AI seat and Microsoft's quiet five percent tax on monthly-billed CSP licenses turn small per-seat changes into meaningful annual numbers, and consumption pricing amplifies that same dynamic across every AI-enabled tool you own. The defensive move is not to slow down on AI, which very few organizations can actually afford to do, but to instrument the spend so that adoption, entitlement, and cost sit in one place before the renewal conversation rather than after it.

How does a vendor-neutral advisor put this to work?

This is where our model matters, and where we want to be precise about the line between the platform and the advisor who works alongside your team. Calero builds and maintains the platform, and we are the vendor-neutral advisor that sits shotgun with your people and turns that visibility into decisions, negotiations, and executed change across all seven layers of your technology estate. We do not resell anything and we are not tied to any single vendor, so the recommendation you receive is grounded in capability and fit rather than in whatever a catalogue happens to pay best. We are deliberately transparent about how that works, because we earn our revenue from the suppliers we represent, it does not change your pricing, and saving you money is how we make money. The platform itself typically runs one to four percent of spend and is backed by a savings guarantee, so if we cannot deliver at least the 20 to 30 percent we average across recurring technology spend, we reimburse the difference and the engagement stays net-zero to you.

What should you do before your next AI renewal?

If you want to get ahead of this before your next true-up, three moves do most of the heavy lifting. First, build a single inventory of every AI and AI-enabled tool in the environment, including the ones embedded in software you already license, because you cannot manage what you cannot see. Second, separate seat-based costs from consumption-based costs and set thresholds and alerts on the consumption side, so a usage spike triggers an early conversation rather than a quarter-end surprise. Third, pull adoption data directly into the renewal itself and reclaim the idle seats and dormant consumption pools that quietly accumulate, which is frequently the fastest dollar you will ever recover. None of these steps requires ripping anything out, because they follow the same sequence we run every day: visibility first, then optimization, then execution.

The bottom line

The headline is simple enough to repeat to your CFO, which is that AI did not create the shadow-spend problem so much as it bolted a meter onto it. The organizations that stay ahead will be the ones that treat AI as one more category inside a single, measurable view of technology spend rather than a separate experiment running off the books. If you want a second set of eyes on what is already hiding inside your stack, that is the conversation we have with mid-market IT and finance teams every week, and it starts with visibility rather than a contract.