Why AI coding agents won’t end the SaaS era

AI coding agents will not eliminate the SaaS business model

AI tools that write code are accelerating development, but they are not rewriting the fundamentals of business. Tom Varsavsky, Chief Technology Officer at SiteMinder, makes this clear: building software has never been the hardest part of building a company. The real challenge is understanding what customers need, earning their trust, and proving consistent value. AI can automate part of the process, but it cannot replace the human insight and operational discipline that keep customers loyal.

Some investors fear that AI will allow every company to recreate complex SaaS platforms on their own. That idea ignores the cost of doing so and the lost opportunity elsewhere. Developing an internal version of a system like Salesforce might be technically possible with AI help, but doing it well would consume resources that are better spent improving customer-focused products. Varsavsky’s team at SiteMinder is using AI to speed up delivery across its existing platform.

Executives should view AI as leverage. It expands capacity and speed, but businesses still win by offering clear value and execution discipline. The SaaS model will evolve, but it will not disappear because its strength lies in customer success.

AI is dramatically accelerating software development productivity

AI is transforming how software gets built. At SiteMinder, adoption of Claude Opus 4.5 has delivered measurable acceleration across the company’s development cycle. Code production is now doubling each month. Work that used to take months finishes within weeks. Smaller teams can complete tasks that previously required much larger groups. Product managers use AI to create prototypes earlier in the process, test ideas faster, and collect customer feedback before moving projects into full engineering.

AI also removes friction from repetitive work. SiteMinder’s engineers now automate large-scale updates and code refactoring, freeing other teams to focus on delivering new customer value. A single “recipe” written by an experienced engineer can manage updates across multiple codebases, cutting time from days to hours. This type of repeatable automation shows how AI helps teams operate at scale without scaling headcount.

For leaders, these improvements demand balance. Speed is a competitive advantage, but it must be matched by strong standards for quality and security. Companies using AI well are not simply faster, they are smarter, with clear systems that measure results at each step. SiteMinder’s experience signals the next phase of SaaS productivity: fewer silos, shorter feedback loops, and more output with the same or fewer resources.

Agentic coding has shifted the workflow from manual programming to AI-driven development

AI is changing the rhythm of software creation. Developers no longer spend most of their time typing code from scratch. Instead, they design outcomes and guide AI agents that generate the code for them. At SiteMinder, this has become the dominant workflow. Tom Varsavsky, the company’s CTO, describes spending only about ten percent of his time reviewing code while the rest is focused on directing the AI agent.

This shift has not removed the need for skilled engineers. It has made their work more strategic. AI handles the repetitive structure of programming, but engineers still define the architecture, test for quality, and ensure reliability. SiteMinder continues to apply code reviews and quality controls across its system, especially because its platform supports hotel bookings that cannot fail. By combining AI-driven speed with established engineering standards, it achieves both fast delivery and stable performance.

For executives, the key takeaway is that agentic coding amplifies human capability rather than replaces it. It demands new management approaches, strong oversight, smart resource allocation, and continuous upskilling. Productivity gains are significant. Internal data at SiteMinder shows that AI-assisted QA processes alone save around thirty minutes per engineer each day. Leadership focus now shifts from managing code detail to aligning teams around AI-enhanced execution.

Managing AI “token economics” will be a critical focus for CTOs moving forward

AI efficiency is not just about speed, it is also about cost. As models such as Claude Opus 4.5 become part of everyday operations, the usage costs of tokens, the units of computation that drive these models, start to grow sharply. Tom Varsavsky views this as the next major management challenge for technology leaders. SiteMinder’s AI usage is doubling monthly, positioning Claude among its top five vendors. At this stage, the priority is effective adoption rather than cost optimization, but Varsavsky expects that focus to shift soon toward measurement and return on value.

He describes token economics as the emerging discipline that will define how AI investments are managed. Just as cloud computing introduced variable cost structures that needed smarter budgeting, AI tokens require continuous monitoring and data-backed evaluation. The ability to balance token consumption with productivity outcomes will separate efficient operators from inefficient ones.

For C-suite leaders, this means preparing for new financial and operational metrics. Tracking productivity in software development has always been complex, but the AI layer adds a new dimension. Clear measurement systems will be needed to assess how each model, team, and workflow contributes to real business output. SiteMinder is already building analytics to understand these differences among teams. The goal is simple: derive maximum productivity per token spent, secure scalable AI usage, and ensure the improvements directly contribute to business growth.

Structured internal governance and cultural alignment are essential for successful AI integration

Widespread AI adoption inside a company doesn’t happen automatically. It requires structure, leadership, and a culture that encourages experimentation. At SiteMinder, Tom Varsavsky, the company’s CTO, has created a clear framework to drive AI integration across teams. This framework includes a developer experience team to test and deploy new tools, a “guild of champions” that shares methods and results, an internal Slack channel for real‑time updates, and a company‑wide AI hackathon designed to encourage participation from every department.

This structured approach ensures that progress isn’t limited to early adopters. Some employees initially resist AI workflows, either out of concern for output quality or a comfort with established routines. SiteMinder addresses this resistance through individual coaching, shared success stories, and evidence of measurable results. Teams that use AI tools report a 50 percent improvement in feature delivery speed, demonstrating how coordinated cultural programs can produce real performance benefits.

For C‑suite leaders, the lesson is that technology adoption scales only when it’s backed by organizational design. Governance is not about restriction, it’s about enabling consistent execution while keeping security and quality intact. As AI changes how technical teams work, maintaining clear standards for architecture, data protection, and development practices becomes a competitive necessity.

Varsavsky’s focus remains on creating a platform that is reliable, secure, and easy to extend. With the right guardrails and alignment, smaller and faster teams can deliver more without additional oversight. Executives who prioritize culture, governance, and measurable progress will build organizations that use AI productively and sustainably, delivering customer value while maintaining control over the technology that powers it.

Key executive takeaways

• AI won’t replace SaaS, it will refine it: SaaS success depends on customer value, trust, and scale, areas where AI cannot substitute human insight. Leaders should invest AI capacity in enhancing existing platforms.
• AI delivers exponential productivity gains: AI tools now double code output monthly at SiteMinder, shrinking project timelines from months to weeks. Executives should leverage similar efficiencies to accelerate delivery while safeguarding quality and security.
• Agentic coding demands new management focus: Developers now direct AI agents rather than write code line by line. Leaders should pivot toward governance, training, and oversight to ensure reliability and accountability in AI-produced software.
• Token economics will shape future cost control: As AI usage scales, consumption-based costs must be tracked and optimized like cloud resources. CTOs should implement clear metrics linking token spend to measurable productivity and business outcomes.
• Structured adoption drives sustainable AI success: Strong governance, shared learning, and cultural alignment enable faster, safer AI integration. Executives should formalize adoption frameworks that balance experimentation with platform stability and security.