Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Simon Mendoza, Chief Technology Officer, Calero
One of my key responsibilities is charting Calero’s technology direction for the years ahead. Right now, a huge part of that is wrestling with how to adopt AI responsibly. An approach that maximizes the benefits of AI for our teams and customers while managing the very real risks around security, quality, and data stewardship is what I’m after.
It’s a fascinating challenge, because there’s no single, settled view of AI across the market today. As a company, we’re deliberately measured in how we approach it: recognizing the potential value of AI while being careful not to move faster than our ability to govern it responsibly. That caution informs how we assess new use cases, set guardrails, and ensure trust, security, and quality remain non-negotiable as we move forward. What strikes me across all these conversations is that most organizations (regardless of whether they’re AI-forward or AI-skeptical) are still figuring out something more fundamental than strategy. They’re struggling with visibility. And I predict that struggle is about to become a crisis.
Here’s my prediction: by the middle of next year, the defining question for enterprise AI won’t be “What’s our AI strategy?” It will be “What AI are we actually using?” And most organizations won’t have a good answer.
Prediction: The First Major “AI via SaaS” Incident Forces a Reckoning
Right now, boards and executive teams are focused on AI roadmaps and transformation initiatives. That’s the exciting conversation. But there’s a problem brewing quietly in the background: AI is already embedded across your technology stack, and most organizations have no visibility into it.
Your CRM added Einstein. Your collaboration platform added Copilot. Your customer service tool added AI routing. Your contract management system added AI extraction. These weren’t procurement decisions, they were feature updates. Your SaaS vendors made the choice for you, and suddenly AI is processing your customer data, your financial information, your strategic documents across dozens of platforms without anyone explicitly deciding to adopt AI in those contexts.
This creates a peculiar problem for both camps. The AI enthusiasts who want to move fast find themselves accidentally running AI they didn’t choose and can’t configure. The AI skeptics who want to move carefully discover their “no AI” policy is already being violated by their existing software stack. Neither group has what they actually need: visibility. Even the most �careful’ teams can be blindsided here.
This isn’t theoretical. At Calero, we’re already tracking AI as a distinct spend category alongside telephony and SaaS, and the growth curve is striking. But the concerning part isn’t the growth – it’s how many of these subscriptions exist outside any formal approval process. Shadow AI makes shadow IT look quaint by comparison.
I predict that by the end of 2026, we’ll see the first major “embedded AI” incident that forces this conversation into the boardroom. It might be a data breach traced back to an embedded AI feature in a SaaS tool. It might be a compliance violation when someone discovers customer data was processed by an unapproved AI tool. It might be a cost overrun when API usage explodes beyond anyone’s projections. Whatever the trigger, it will expose a fundamental truth: you can’t manage what you can’t see.
Prediction: Audit Committees Will Demand AI Visibility
When that incident hits, the audit committee response will be predictable: “Show me our AI inventory. What tools do we have? Where is our data going? What are we spending? What’s our exposure?”
And for most organizations, the honest answer will be: “We don’t know.”
This is where 2026 is going to get interesting. Organizations will scramble to create visibility, and they’ll quickly realize there are two very different problems here. There’s the highly technical challenge of governing custom AI infrastructure, proprietary models, and complex data pipelines. This requires deep ML expertise and sophisticated tooling. But there’s also the challenge of discovering and managing subscription-based AI: the SaaS platforms with embedded AI features, the standalone AI tool subscriptions, and the license-based services that employees are adopting across departments.
That second category represents the vast majority of enterprise AI exposure, and it’s entirely manageable with existing approaches. The same discovery methods that find shadow SaaS can find shadow AI. The same vendor management frameworks that govern SaaS contracts can govern AI subscriptions. The same optimization playbook that eliminates redundant software can eliminate redundant AI tools.
The organizations that will handle this crisis smoothly are the ones that already invested in a mature technology expense management program. They have the discovery infrastructure. They have the telemetry. They have the vendor relationships and contract visibility. They understand that managing technology spend requires seeing it first, controlling it second, and optimizing it third. You can’t skip the first step.
Prediction: The Visibility Gap Will Separate the Prepared from the Scrambling
At Calero, we’ve been managing recurring technology expenses for mid-market and enterprise organizations for over two decades. We watched telephony and network sprawl happen as new technologies supplanted old. We watched mobility explode, with minimal governance, as it unlocked incredible productivity gains. We are watching SaaS sprawl happen right now. The pattern is always the same: initial adoption is decentralized and enthusiastic, visibility lags behind usage, costs balloon, risks accumulate, and eventually someone has to bring order to chaos. AI is following the same script, just faster and with higher stakes.
The difference in 2026 won’t be between organizations that have AI and those that don’t: everyone will have AI, whether they planned for it or not. The difference will be between those who can govern it and those who can’t.
Visibility isn’t a competitive advantage, it’s table stakes for responsible operation. But when the incident hits, when the audit committee asks uncomfortable questions, when the scramble begins, not having it becomes a serious disadvantage. The organizations with mature technology expense management won’t be heroes. They’ll just be ready. Everyone else will be explaining why they weren’t.
You can’t manage what you can’t see. In 2026, the value of thoughtful preparation will become clear.
##
ABOUT THE AUTHOR
Simon Mendoza is Chief Technology Officer at Calero, focusing on the intersection of AI adoption, enterprise governance, and technology spend visibility. He brings a practitioner’s perspective built from two decades architecting and scaling enterprise platforms – now applied to the challenge of managing AI exposure across complex software estates and helping organizations build the infrastructure to see, control, and optimize technology investments before visibility gaps become crises.
