For 25 years, the SaaS model promised to simplify business. Instead, it delivered escalating costs, rigid workflows and expensive implementation cycles. Through “vibe coding”—a radically streamlined process by which to build software solutions with the help of agentic AI—we are moving from a world in which we rent a solution to a world in which we describe our business problems to an AI agent and the technology builds itself around us.

Listen and Learn “How Agentic AI is Killing the SaaS Seat License” Deep Dive

The Genie is Out of The Lamp

We’ve all heard the folk tale of Aladdin and The Wonderful Lamp, in which Aladdin is granted three wishes by the lamp’s resident genie. In some respects, we all now have access to a genie to help us solve some pretty real problems—especially when “wishing for technology that can do X.”

AI coding tools and agents from Anthropic, OpenAI and Google are granting these wishes by allowing us to verbally describe the features, functionality and desired outcomes we need. For a $20–$50 monthly subscription and a few hours of iteration, the “wish” becomes a working reality.

From Months to Minutes: AI-Powered Development Is Already Here

Consider this phenomenal shift in speed: one developer recently collapsed two months of work into two hours. In 2015, it took him two months of learning, iteration and debugging to build a native Android-only application. In 2025, he used vibe coding to rebuild it as a polished, cross-platform progressive web app in just two hours.

Or take this enterprise example: the company Anything, a vibe-coded AI app platform that launched last September, provides the tools—from databases to storage and payment functionality—that users need to run businesses on the web. Its initial traction was explosive, reaching a $2 million annualized run rate in just two weeks.

The $428 Billion Question: What Happens to SaaS?

By now, we’re all familiar with how AI will displace jobs. But what about how AI-powered coding will impact the $428.78B global SaaS market, where in 2024 companies spent an average of $8,700 per employee across an average of 106 SaaS apps?

When Salesforce started in 1999, it was built on the premise of having companies “stop buying software and instead start renting it.” Yet, what started as a way to simplify and reduce costs has evolved into the opposite. Gartner estimates that 30% of all SaaS spend is “toxic,” i.e., money spent on unused licenses and features that deliver zero value. Furthermore, seat-based pricing models are fundamentally broken when work shifts from human employees to AI agents.

“The real challenge of SaaS was never just the cost. It was the compromise.”

The market is already reacting. On February 3rd, approximately $300 billion in market value was wiped out in just 24 hours, affecting giants like Salesforce, ServiceNow and HubSpot. Valuation multiples plummeted as OpenAI and Anthropic debuted agentic platforms that allow companies to build and manage their own custom technology layers. In this new world, the prompt is the platform.

The real challenge of SaaS was never just the cost. It was the compromise. Every company that buys a SaaS license is essentially forcing their unique business processes into someone else’s box. When it doesn’t fit, they hire costly system integrators to build workarounds, bolt on third-party apps and spend years maintaining a solution that’s never quite right.

AI changes this equation entirely. For the first time, a business leader can describe the exact problem they need to solve and get a tool built around that specific wish—rather than renting a generic tool and hoping it’s “close enough” to solve the problem.

The Health Plan Shift: A Case Study

Ten years ago, I was with a health insurance technology startup as their Chief Brand and Digital Marketing Officer. We were building a SaaS platform to give payers a “fishbowl” view into their members. The goal was to understand which members were the most profitable versus those that were costing the health plan money—e.g., members who weren’t managing chronic conditions or were driving up costs through preventable ER visits.

The concept was powerful and it was filling an industry void at the time, but like most enterprise SaaS at the time, the path from purchase to value was long and complex. The health plan would buy the platform, then most likely invest in system integrators to align their member data and plan structures with the platform’s architecture. They’d pay per-seat licenses for every user, then wait months for full implementation.

Through the lens of agentic AI, that paradigm is dead. The “integration” is no longer a manual labor project; rather, it’s a linguistic instruction. Instead of forcing their unique data into someone else’s rigid structure, the team simply vibe codes their requirements into existence.

For example, they might issue the following instructions to the agentic AI:

“I need to predict which members will become high-cost before they actually hit that threshold. I want this built on a compliant architecture using our own claims data and real-time health indicators. Give me automated alerts to department heads the moment we trend toward a cost spike. Every Monday, I need action plans with specific interventions delivered to the right stakeholders, plus real-time optimization tweaks. Most importantly, I want a visual dashboard that tells us exactly how we’re tracking against our business goals.”

Static and Working Demo Dashboard

This isn’t an 18-month implementation with a $2 million price tag. It happens in weeks, built around their data, their plan structure, and their specific cost drivers. No “toxic” seat licenses. No system integrators. No forcing a unique business problem into a generic box. In fact, the static dashboard below took just a few minutes for Claude to render, and the interactive version you can view here took about 20 minutes for Claude to code and for me to upload to my server.

The vibe coding solution is fundamentally better: it’s proactive instead of reactive and predictive instead of retrospective. The technology didn’t just get faster — it got smarter, and so did the possibilities.

The Strategic Warning

The companies that thrive in this reality will be those that deeply understand their own problems. For them, AI is transformational.
But for companies without clarity? They’ll just build the wrong solutions faster and cheaper. Speed only makes a bad strategy more expensive. If you don’t have a clear rationale, you’ll just spin up dozens of half-baked tools that don’t solve the core issue.

It is no longer about who can build the technology; rather, it’s about who knows what to build. That’s a strategy and insight problem, not a technology problem. The question every CMO, PE partner, and board director should be asking right now is: Do we understand our business well enough to know what solution we’re actually wishing for?

For more on AI strategy and its business impact, visit scheinerinc.com/perspectives.

To continue the conversation, you can reach me at mike@scheinerinc.com.

As always, thank you for reading and/or listening.