Many of the AI companies we work with are building highly differentiated products. But behind the product, we often see finance functions still structured around traditional SaaS assumptions.
That gap creates real challenges, particularly as companies scale. Revenue recognition becomes inconsistent, metrics lose meaning, margins become harder to interpret, and investor conversations become more challenging.
Token-based, consumption-driven revenue is not simply a pricing shift. It introduces a different financial model, one that requires a more integrated approach to accounting, forecasting, and decision-making.
For founders, understanding this shift early is critical to building a durable, scalable business.
From Subscription to Consumption
Traditional SaaS models are designed around predictability. Revenue is contracted, usage is secondary, and recognition is typically ratable over time.
AI pricing models are different. They are consumption-based:
- Revenue is tied to usage
- Usage can fluctuate significantly
- Costs are incurred in real time with each interaction
That shift introduces variability into systems that were historically designed for stability. As a result, finance needs to move closer to operations—capturing how revenue is actually generated, not just how it is contracted.
Revenue Recognition in a Consumption Model
Under ASC 606, revenue is recognized as tokens are consumed, not over a fixed subscription period. While straightforward in principle, this creates complexity in practice.
Three areas require particular attention:
Prepaid Credits
Upfront purchases of credits represent deferred revenue, not recognized income. Revenue is earned only as usage occurs. Clear policies around expiration and breakage are essential, both for audit readiness and consistency in reporting.
Multi-Element Arrangements
Many AI products bundle platform access with usage-based pricing. Allocating revenue across these elements requires a Standalone Selling Price (SSP) analysis. This step is often overlooked early but becomes increasingly important as companies approach audit or diligence.
Cost Alignment
Compute and infrastructure costs are incurred at the point of usage. If revenue recognition is not aligned with these costs, gross margin can appear inconsistent across periods, making performance harder to interpret for both management and investors.
Rethinking Core Metrics
Metrics that worked well for SaaS do not always translate cleanly to consumption-based models.
For example, ARR can become less informative when usage varies month to month. Instead, companies benefit from incorporating metrics that reflect actual behavior and economics:
- Revenue Run rate with volatility bands rather than ARR
- Active consumption users instead of seat count
- Usage expansion by cohort in place of traditional NRR
- Average consumption value (trailing 90 days) in place of ACV
In addition, several metrics become more important:
- Fully Loaded Token Margin – A comprehensive view of unit economics, incorporating all costs associated with delivering each unit of usage.
- Consumption Velocity – The rate at which customers adopt and expand usage over time is often an early indicator of retention.
- Usage Concentration – The degree to which revenue is concentrated among a small number of users can introduce churn risk.
These metrics provide a more accurate picture of performance in a usage-driven environment.
Forecasting in a Variable Environment
Forecasting also requires a different approach.
Rather than relying on contracted revenue alone, top performing clients model usage behavior:
- Segment customers by how they actually consume the product
- Incorporate seasonal or event-driven usage patterns
- Use committed minimums as a baseline where applicable
- Build scenario-based forecasts to reflect variability
The goal is not to eliminate uncertainty but to understand and plan for it.
Margins as a Cross-Functional Lever
In consumption-based models, margin is influenced by both pricing and technical decisions.
One of the most effective approaches we see is a close partnership between finance and engineering:
- Align on cost structures at the model or feature level
- Route usage to the appropriate level of compute based on task complexity
- Continuously evaluate tradeoffs between cost, performance, and customer experience
This type of coordination can materially improve unit economics and create more flexibility in pricing and growth strategy.
It is also important to recognize that early-stage margins may not reflect long-term performance. As usage scales, cost structures often improve. Modeling that trajectory helps avoid premature decisions based on early data.
Building the Right Foundation Early
A consistent theme across successful companies is that financial infrastructure is built alongside product infrastructure—not after the fact.
At a minimum, this includes:
- A detailed usage ledger – Tracking usage at a granular level (customer, timing, model, cost) to support accurate revenue recognition and cost analysis.
- Systems aligned to consumption models – Ensuring billing, CRM, and reporting tools reflect how revenue is actually generated.
- Early visibility into unit economics – Establishing dashboards and reporting that connect usage to financial outcomes.
- Documented accounting policies – Defining how revenue, credits, and variable consideration are treated, and applying those policies consistently.
Investing in this foundation early reduces complexity later, particularly during fundraising or audit processes.
A More Integrated Role for Finance
As AI companies scale, finance plays a more integrated role in the business:
- Connecting product usage to financial outcomes
- Supporting decision-making with real-time insights
- Ensuring financial data reflects operational reality
Companies that approach finance in this way are better positioned to communicate clearly with investors and make more informed strategic decisions.
How Attivo Supports This Transition
At Attivo, we work with venture-backed companies navigating the shift to consumption-based models.
Our focus is on helping teams:
- Build finance and accounting systems that reflect usage-driven revenue
- Develop forecasting models that incorporate variability and real cost structures
- Establish clear, consistent policies that support audit and diligence readiness
- Translate complex financial dynamics into actionable insights
In this environment, finance is not just a reporting function; it is a core part of how companies scale effectively.
