The Great Relayering: SaaS Vendors Rework Pricing As AI Changes Cost Structures

Generative AI has shifted the baseline for the feature sets enterprise software vendors spent years developing, while introducing a dynamic cost structure that traditional SaaS margins are now adapting to absorb. This shift represents a practical economic relayering: a period of recalibration where vendors are optimizing their pricing and delivery models to align with the concrete, high-impact results that today’s buyers are expecting.

Joseph Daee is a Chief Revenue Officer and VP of Enterprise Sales with more than two decades of experience managing high-stakes P&Ls and complex enterprise deals. Over the course of five companies, he oversaw $279 million in revenue expansion and reduced customer acquisition costs by 89 percent during his tenure as VP of Enterprise Sales at Lofty, an AI-infused CRM platform. He sees the current market shift as a practical math problem with very real consequences for how revenue teams operate.

"AI isn't killing SaaS. It's forcing a relayering of how we price, sell, and deliver value. The real shift is moving away from access and toward outcomes that customers can measure," says Daee. This shift is driven by a two-sided market dynamic. On one side, technically proficient buyers are leveraging large language models to build their own lightweight tools. On the other, vendors are managing backend compute costs that scale with usage, breaking the comfortable margin structure most SaaS businesses were built on. As core features become easier to replicate and harder to monetize, enterprise buyers are now focused on the specific, high-level value that justifies their investment.

• Governance as the moat: Daee tested the DIY impulse firsthand, building a mini CRM inside Claude to explore its potential. The experiment reinforced the distinction between basic accessibility and true enterprise readiness. "Small companies building homegrown tools often lack the time or resources to match the specialized development required for modern SaaS platforms," he notes. "In its current state, AI alone doesn't provide the full security and governance required to deliver a solution that is consistently safe and reliable for the end user."

For vendors seeing feature-level commoditization accelerate, security and compliance serve as a tactical baseline against churn. The question for buyers shifts from whether they can build a tool to whether they should, considering the operational risk of maintaining it without dedicated infrastructure, continuous iteration, and enterprise-grade data protection.

• End-to-end integration: Significant efficiency gains from AI result from layering automation across the entire workflow rather than relying on isolated tools. Daee sees companies moving toward deeper, end-to-end AI integration within established platforms rather than bolting individual features on top. "It’s about layering AI from beginning to end across the entire work and automation process," he says. For large enterprises with complex tech stacks, achieving that level of integration requires the kind of sustained investment that professional platforms provide and homegrown solutions rarely match.

That governance layer sits on top of a cost structure that is evolving away from traditional SaaS models. While software companies historically enjoyed 80% to 90% margins, AI token compute introduces a more volatile cost of goods sold that fluctuates with adoption and usage. This shift is making it difficult for vendors to price their products with the same predictability. Rising fuel and energy costs compound the problem further, making data centers even more expensive to operate and putting additional pressure on vendors still searching for a sustainable pricing floor.

• The token toll: In a recent VP of Enterprise Sales position, Daee watched these cost pressures play out in real time. The company initially bundled AI features into the core product at no additional cost before backend expenses prompted a strategic pricing rethink. "The cost of AI usage became a significant factor that shifted our SaaS margins," he recalls. "We moved to implement token tiers, which has since become the industry standard."

• Shedding idle seats: The cost pressure coincides with a buyer-side reckoning that predates the AI boom. Many enterprise organizations have spent years paying for idle licenses, buying hundreds of seats when only a fraction are actively used. "Companies might buy 500 seats, but if those seats aren’t being fully utilized, it represents an opportunity to optimize their spend,” Daee observes. At a previous company, his team successfully transitioned toward usage-based pricing, a strategic shift he sees the broader market heading.

• Monetization lever: Finite compute capacity also creates a unique market opportunity. Daee points to a recent weekend when a massive wave of user migration from one AI platform to another caused significant server latency. "For those seeking priority status, we can offer guaranteed access and zero delays for a premium," Daee explains. This shift mirrors entertainment industry pricing, where tiered subscriptions separate basic and premium experiences, suggesting that software vendors will compete on reliability and speed alongside traditional features.

• Floating on funding: Much of the current experimentation is supported by venture capital, which subsidizes pricing and allows vendors to test new models before profitability becomes a primary requirement. "As long as VC capital remains available, the market will continue to evolve gradually," Daee says. "But we are reaching a point where we’re bringing back old-school managed services and integrating them into this new foundational AI SaaS market." As the venture buffer narrows, vendors are increasingly focused on aligning their pricing with real cost structures to ensure long-term stability.

As pricing aligns with outcomes, the sales motion is evolving to match. Revenue teams are moving beyond theoretical ROI spreadsheets and platform demonstrations to provide impact that buyers can verify independently. This shift is bringing enterprise selling closer to a high-touch, results-oriented model than traditional software licensing.

• Skin in the game: Some platform leaders are exploring flat fees paired with performance-based contingencies, a strategy borrowed from proven commercial models. “Historically, managed services were often paid per engagement or on a contingency basis, where saving a client a million dollars earned a 25 percent share of that success," Daee says. "Vendors are reintroducing these legacy models to ensure AI costs are transparent and value-aligned for the customer."

• Proof over promises: Enterprise buyers are starting to view AI investments as strategic assets rather than speculative bets. They prioritize hard numbers and documented results from comparable organizations before finalizing contracts. "They value references, white papers, and case studies that provide the specific details of how a solution was used, how it impacted the business, and the concrete results it delivered."

While buyers want accountability, they also seek partners who can navigate the complexities of onboarding and implementation at scale. The vendors best positioned to thrive are those that balance credible measurement with the kind of high-touch customer engagement that builds long-term trust. As features become more accessible and backend costs grow, success will depend on more than just the software itself. The vendors navigating this shift most effectively are those that deliver sound governance, align pricing with verified outcomes, and bring documented proof to every conversation. "People want accountability," Daee concludes. "They want to know what they're going to get and what they're going to pay."