SaaS Founders Must Prove ROI, Not Just Growth, as AI Reshapes Valuations

The SaaS playbook is rewriting in real time

For three decades, SaaS metrics were simple and sticky: predictable revenue, high gross margins, efficient customer acquisition, and strong net revenue retention defined venture returns. Today, LLM commoditization and margin compression are forcing a reset (per Ivan Nikkhoo, managing partner at Navigate Ventures).

Founders are hearing conflicting advice. Sequoia partner Julien Bek recently argued that the next trillion-dollar company will be "software disguised as a services firm," capturing services spend alongside software revenue. The logic: for every dollar spent on software, six are spent on services. But Nikkhoo warns this framing conflates two separate investor concerns—portfolio risk and model economics—into a single pivot recommendation that most founders should ignore.

The actual market reset is narrower and harder: investors are now judging SaaS on capital efficiency, gross and net retention, Rule of 40, CAC payback, and burn multiple. Hypergrowth alone no longer attracts capital if retention is unproven or switching costs are low. A demo is not enough.

Defensibility, not novelty, determines who survives the AI wave

AI startups can grow at unprecedented rates in their first 12 months, but that early velocity masks structural weakness if the product lacks durable moats. Promising AI categories are attracting 2x to 3x more competitors than prior waves (company-reported), and large SaaS incumbents—Salesforce, ServiceNow, Microsoft—are shipping AI features faster than startups can prove theirs matter.

The bar for "what matters" has shifted from software access (seats, users, departments) to workflow ownership and measurable outcomes. If your AI performs work independently, seat-based pricing breaks. Pricing is migrating to usage-, consumption-, and outcome-based models (per Battery Ventures' State of AI Report). Investors will ask directly: Is this a defensible position or a feature a giant can ship in a quarter?

For founders, the implication is discipline over trend-chasing. Domain expertise in customer workflows, clarity on when to use frontier models versus smaller specialized models, and economic tradeoffs around fine-tuning and human review are now table stakes. Companies that understand their customer's pain with evidence of urgent demand, prove retention, and show efficiency metrics under scrutiny will get funded. Generalist AI plays will not.

What founders should actually do

Stop listening to advisory noise about layering services on top of software unless that bundled offering is core to your customer's operations. If it is not, a pivot will likely be necessary—but only after you have proven the software moat is real.

Build a sharp wedge into a specific buyer segment, demonstrate strong product usage and measurable ROI, and map a clear roadmap from point solution to platform. Your pitch must answer this: Can this company grow efficiently and organically, retain customers through budget scrutiny, and compound value as it scales? The answer lives in your metrics, not your market narrative.

Pricing should reflect how your product delivers value. If AI is executing work directly, shift away from seats. If your moat depends on domain data or workflow integration, price on outcomes achieved or consumption. If you are still using frontier models for every inference, cost-of-intelligence arithmetic will kill your margins.

Finally, understand that AI expands the addressable market significantly beyond cloud software into services automation and labor displacement. That is the real opportunity. Founders who exploit it will be operators who can map that expansion to their specific customer's workflow, not those who chase every AI trend that a VC blog post glamorizes.