The landscape of enterprise technology is undergoing a fundamental restructuring, moving away from the decades-old reliance on user seats toward a more dynamic, performance-based architecture. As a specialist in enterprise SaaS design and architecture, Vijay Raina has been at the forefront of this transition, helping organizations navigate the complexities of AI integration and the economic shifts it triggers. With a deep understanding of how software adds value to the modern workforce, Raina offers a unique perspective on why traditional metrics are failing and how “credit-centric” models are becoming the new gold standard for the industry.
This discussion explores the five major forces driving the shift toward consumption-based value, the collapse of implementation timelines, and the emergence of specialized financial metrics like “Inference-adjusted gross margin.” Raina explains how the shift from software budgets to labor budgets is expanding the total market opportunity by up to ten times while simultaneously forcing a “bifurcation” of customer retention where the middle ground is rapidly disappearing.
The traditional SaaS model was built on the concept of “seats,” yet we are seeing a massive shift toward credit-centric metrics. Why is the seat-based model no longer sufficient for the AI-native era?
The fundamental unit of value has moved from who has access to the software to what the software actually achieves. In this new AI-native world, we aren’t just selling a login to a human employee; we are deploying autonomous agents that handle entire workloads, and those agents don’t occupy physical “seats” in the traditional sense. Over the past 18 months, we have observed every major SaaS platform pivot toward some form of credit or consumption unit because the cost of goods sold is exploding due to AI inference. In the old days, a software company could scale to an infinite number of users with nearly zero marginal cost, but today, every time a user triggers an AI action, there is a tangible cost for compute that scales linearly with usage. This shift forces a reality check on the P&L statement, moving the conversation away from simple access and toward a rigorous measurement of work performed.
As purchasing decisions move higher up the organizational chart to COOs and line-of-business leaders, how is this changing the total addressable market for software vendors?
When you move the conversation from software budgets to labor budgets, the ceiling for growth shatters completely. We are seeing a shift where AI-native applications are no longer just tools for employees, but are actually replacing or augmenting professional services and manual labor, which expands the total market opportunity by 3x to 10x compared to traditional SaaS. Business leaders are no longer looking for a better spreadsheet; they are looking for an agent that can recover revenue or resolve tickets autonomously. This expansion into the labor pool means the stakes are much higher, and the software is being judged on its ability to drive operational outcomes rather than its feature list. It’s a profound moment where software starts to act like a digital workforce rather than a static utility.
Pricing is becoming increasingly variable, ranging from tokens to outcomes. Where do you see most defensible AI businesses landing on this spectrum?
While subscription pricing offers stability and tokens offer transparency for commoditized features, “credits” are rapidly becoming the dominant architecture for defensible businesses. Credits provide a clever buffer for the vendor, allowing them to set a conversion ratio that protects their margins while shielding the customer from the granular, often confusing volatility of inference costs. This model allows for predictable annual contracts that can still account for high-usage peaks across multiple product lines. Some verticals are pushing toward outcome-based pricing—charging for a specific result like a qualified lead—but that is currently limited to niches where the AI’s impact is indisputable and easily tracked. For most of the industry, credits strike the perfect balance between customer predictability and the vendor’s need to control the margin of their AI workloads.
You’ve mentioned that “Time to Value” is collapsing in the AI era. What does this mean for the traditional customer success and onboarding journey?
The era of implementation cycles lasting several quarters is effectively over, as AI-native tools are designed to deliver meaningful results in weeks rather than months. We are seeing a collapse in onboarding friction because workflows are increasingly pre-built and agents can be deployed into existing data environments with minimal manual configuration. This speed eliminates the need for the heavy, human-intensive professional services that used to define enterprise software rollouts. From a sensory perspective, the “wow” moment for the customer happens almost instantly, which changes the entire psychological contract between the buyer and the vendor. If a tool doesn’t prove its worth within that first month, it is likely to be discarded, which is why we are seeing such a sharp bifurcation in retention rates.
How should finance leaders rethink their metrics, specifically regarding revenue composition and gross margins, to account for these new AI realities?
We have to move beyond “blended” SaaS margins and start looking at “Inference-adjusted gross margins” to see the true economic health of an AI-native company. One of the most critical new metrics is the “committed credit ARR” versus “burndown ARR,” because this tells investors if your revenue is behaving like a stable subscription or a volatile usage model. We are also tracking “credit burn velocity” to see how fast customers are consuming their prepaid units, which acts as a leading indicator of expansion long before the renewal date hits. To keep the business healthy, the “AI-adjusted Rule of 40” is being used to recalibrate growth and profitability targets, accounting for the reality that AI margins might be lower than the 80% or 90% gross margins we saw in the previous decade. It’s about creating a clearer picture of how efficiently a company converts its compute spend into actual revenue.
With the rise of “credit margin engineering,” how is the internal operating model of a software company shifting?
We are witnessing an inversion of the classic operating model where R&D and the cost of goods sold are moving up the P&L, while sales, marketing, and customer success are moving down. In this new structure, the software itself does more of the “selling” and “retaining” through automated outcomes, which reduces the need for massive human-centric success teams. The most successful operators are now focusing on “credit margin engineering,” which is the deliberate optimization of the gap between what a credit costs the customer and what it costs the vendor to execute the underlying compute. Those who can scale their compute leverage ratio—getting more revenue out of every dollar spent on inference—will be the ones who dominate the market. It’s a more technical, architecturally-driven way to run a business, where the efficiency of the code directly dictates the health of the balance sheet.
What is your forecast for the future of enterprise SaaS valuations as these new credit-centric metrics become standard?
In the coming years, we will see a massive divergence in valuations based on the “commit-to-burndown” ratio of a company’s revenue. A business that can show 80% of its credits are committed through annual contracts will trade at high subscription multiples, while those relying on volatile, month-to-month usage will be valued more like utilities or service providers. Investors will stop looking at “seats” entirely and will instead obsess over “credit utilization rates” as the ultimate predictor of churn. If a customer isn’t burning through their credits, a downsize at renewal is virtually guaranteed, making utilization the new “North Star” for growth. Ultimately, the companies that thrive will be those that can prove their AI agents are performing high-value work, effectively turning software into a high-margin digital labor force.
