May 2026 Summaries

As AI and usage-based pricing models become prevalent, traditional software licenses, which merely grant the legal right to use software, are insufficient for determining real-time access and capabilities within a product. License entitlements emerge as a dynamic alternative, allowing for real-time evaluation of customer access based on active plans, usage, and various overrides, which are crucial for AI products that consume compute resources per request. These entitlements function as a runtime layer, enforcing access policy by continuously evaluating plan configurations and usage states to decide request outcomes. Unlike static licenses that struggle with dynamic pricing and concurrent usage, entitlement systems are built to handle complex scenarios such as tiered plans, add-ons, trials, and promotions, ensuring flexibility and real-time adaptability. By centralizing entitlement logic, engineering teams can avoid scattering access controls across services, mitigating the risk of overspending and inconsistent access decisions under high-traffic AI workloads.

Metronome, a usage-based billing platform acquired by Stripe in January 2026, offers pricing plans that are primarily usage-based with a free Starter tier and custom plans for larger operations. The Starter plan provides engineering teams with tools for event ingestion, usage-based pricing models, and integration with Stripe, making it suitable for those building billing systems without starting from scratch. The Custom plan caters to teams managing complex billing flows across multiple channels, offering additional integrations with systems like Salesforce and NetSuite, and features tailored pricing based on event volume. Metronome's pricing is heavily influenced by how events are designed and aggregated within a system, and while it effectively processes billing data post-event, it does not handle real-time governance or feature access, which is where solutions like Stigg complement it by managing entitlements and credits during the request flow.

Stigg has introduced two new interfaces to enhance its monetization infrastructure: a native MCP server and Stigg CLI, designed to integrate seamlessly with AI coding assistants like Claude and Cursor, allowing developers to manage Stigg through natural language or direct commands without leaving their work environment. The MCP server facilitates tasks such as pricing modeling, credit allocation, and entitlement management using natural language instructions, optimizing the process for AI tools. Conversely, the CLI provides explicit, deterministic commands for operations requiring precise control, such as data migrations and deployment rollouts, catering to teams needing exact API interaction. These advancements reflect a strategic shift towards embedding Stigg in developers' existing environments, prioritizing seamless integration over standalone operation. Both tools are currently available in public beta, inviting developers to incorporate them into their workflows for efficient and predictable control over Stigg's capabilities.

Hybrid pricing models, which combine a fixed subscription fee with variable usage-based charges, present complex infrastructure challenges that many teams do not initially anticipate. This model requires real-time tracking and enforcement of customer entitlements, accurate data transmission to billing systems, and seamless integration between various internal systems to avoid issues such as incorrect billing or enforcement failures. As companies like New Relic have shown, transitioning to hybrid pricing can necessitate significant testing and system adjustments to handle edge cases and ensure scalability. The implementation of hybrid pricing highlights the need for sophisticated metering and entitlement systems to manage real-time cost control, especially for AI features where resource consumption can be unpredictable. Companies often face a decision between building in-house solutions, which may struggle with scalability and complexity over time, or purchasing dedicated infrastructure solutions like Stigg, which offer integrated entitlement enforcement, real-time metering, and credit management. These systems ensure consistent alignment across entitlements, billing, and usage, crucial for maintaining operational efficiency and supporting advanced pricing models.

Metronome is a usage-based billing software tailored for engineering teams, particularly those in AI and SaaS sectors, requiring a metering engine for consumption-based pricing models. It excels in providing SQL-based metrics, real-time event ingestion, and cloud marketplace integrations, making it suitable for teams needing detailed billing logic control and synchronization across platforms like AWS, Azure, and GCP. However, it lacks runtime enforcement, event backfills, and comprehensive credit management, necessitating additional layers for entitlement enforcement. Acquired by Stripe in January 2026, Metronome's integration capabilities are evolving, though it remains more effective when billing is managed by engineering teams. While praised for handling complex billing primitives, users have noted that configuration requires significant time and engineering involvement, particularly for pricing changes and integration enhancements. Despite its strengths, Metronome's reliance on a streaming aggregation model limits retroactive adjustments and keeps pricing tied to engineering workflows, potentially slowing down iterations as teams scale and diversify.

A token in AI refers to the smallest unit of data processed by language models, affecting pricing, usage tracking, and system enforcement in AI applications. Tokenization involves converting text into numeric IDs, with variations in token count depending on the model and input, impacting how usage is measured and billed. AI pricing models often rely on tokens, and these models can be structured in various ways, such as pay-per-token, prepaid credits, subscription with limits, or hybrid models, each requiring specific infrastructure for real-time metering and enforcement. Token metering infrastructure must handle event attribution, real-time enforcement, concurrent session handling, and caching to ensure accurate usage tracking and billing. Entitlements define usage limits based on customer plans, and credits track and limit token usage, requiring a robust system to handle issuance, consumption, and reconciliation. Effective AI usage governance is crucial for managing token consumption across users and teams, necessitating an enforcement layer to make real-time decisions before costs are incurred. As AI systems grow, teams often face challenges with enforcement, shared usage, and plan changes, prompting the need for dedicated runtime infrastructure to bridge the gap between usage tracking and billing systems.

AI credits, increasingly used in AI products, present a complex infrastructure challenge rather than merely a data model issue, necessitating real-time deduction, shared pools, and low-latency enforcement for effective implementation. As AI products evolved, the complexity of managing individual feature pricing led to the adoption of a unified credit system, where one currency is consumed across multiple features, such as Stability AI's use of credits across various capabilities. The choice between subscription-based and prepaid credit models impacts billing, enforcement, and system complexity, with each model suiting different usage predictability and customer commitment levels. Essential components of a credit system include credit wallets, grants, consumption rates, depletion behavior, and rollover policies, all requiring precise engineering to ensure consistent real-time enforcement and financial alignment. Credits must be tracked meticulously, incorporating effective and expiration dates, cost basis, and burn priority, to maintain financial auditability and avoid issues like double deductions or revenue leakage. The infrastructure must support dynamic burn order recalculations and real-time enforcement across multiple features, aligning product behavior with billing and financial records. Comprehensive observability and customer-facing visibility are crucial to maintaining trust and operational reliability, with tools needed for audit logs, grant tracking, and manual overrides. Building such a system from scratch can become an ongoing maintenance burden, prompting some companies to seek specialized solutions like Stigg, which offers a runtime layer for credit management, ensuring product teams can focus on development rather than billing logic.

Pricing strategies in the SaaS industry are evolving rapidly, with companies frequently updating their packaging and pricing structures more often than launching new products. The traditional seat-based model is being challenged by the rise of usage-based pricing, although many companies still retain seats while incorporating AI features as add-ons or bundling them into core plans. Credits are increasingly used on top of existing pricing models, though communicating their usage to customers remains complex. Transparency in pricing is becoming crucial, not just as a customer-friendly practice but as an operational necessity, especially as AI agents begin to play roles in purchasing decisions. The ongoing shift requires companies to innovate continuously, treating pricing as a dynamic process rather than an annual task. Those adapting quickly are integrating seats and consumption models, designing user-centric credit systems, and investing in agile infrastructure to accommodate strategy shifts, while others lag behind by maintaining outdated practices. This evolution underscores the need for a dedicated role at the intersection of monetization and billing infrastructure, mirroring the emergence of GTM engineering roles in recent years.

Monetizing AI products effectively hinges on selecting the right pricing model, enforcing it through robust infrastructure, and adapting swiftly to changing usage patterns. Traditional SaaS infrastructures struggle with AI's real-time usage demands, necessitating systems that can dynamically track and enforce limits on requests to prevent costly overages, as demonstrated by cases like Segment8's costly integration bug. Companies often choose between seat-based, usage-based, or hybrid pricing models, each with distinct implications for margins and infrastructure complexity. Stigg provides a solution with its real-time decision engine, managing entitlements, credits, and usage limits directly within the product, thus enabling swift pricing model changes and effective usage governance without the heavy engineering burden typically required. The infrastructure, including a product catalog and a metering system, allows for efficient enforcement of quotas and credit management, ensuring that monetization efforts can scale alongside business needs.

An entitlement management system (EMS) is crucial for AI companies to efficiently handle customer entitlements and prevent billing overages by defining, administering, and enforcing what each customer can do within a product. Unlike Role-Based Access Control (RBAC), which is binary and focused on internal permissions, an EMS considers commercial allowances from various sources, including plans, add-ons, and promotional grants, to determine customer entitlements. It operates alongside billing systems, such as Stripe or Chargebee, without replacing them, ensuring that credit balances are checked before any billable actions occur. Core components of an EMS include a centralized product catalog, real-time metering, provisioning, feature gating, and enforcement, which together manage access limits and usage in a synchronized manner. Legacy EMS solutions, initially designed for on-premises software licensing, differ from modern EMS tools, which cater to cloud-native SaaS and AI products with real-time entitlement enforcement and complex pricing models. A dedicated EMS becomes essential when entitlements surpass engineering capacity, enabling seamless integration with billing systems and facilitating swift product packaging changes without engineering involvement.

During recent visits to the Bay Area for industry events, conversations with engineering and platform leaders revealed a growing focus on entitlements infrastructure, driven by the demands of SOX compliance, especially as companies prepare for IPOs or face scrutiny from late-stage investors. While traditionally seen as a finance issue, SOX requires technical evidence that customers receive exactly what they paid for, which turns entitlements into an engineering problem. This is particularly challenging for AI companies with dynamic, consumption-based pricing models, as they must provide auditable proof of fulfillment and manage complex credit-based systems. Unlike traditional SaaS models with relatively static entitlements, AI products require real-time tracking of usage and entitlements, which complicates compliance. Public companies like OpenAI and Anthropic are preparing for IPOs with credit-based models, highlighting the need for robust, SOX-compliant infrastructure that can handle high-scale, consumption-based revenue recognition. This shift isn't limited to AI-native companies; established SaaS platforms integrating AI features must also navigate these complexities. Developing a SOX-ready entitlements infrastructure entails maintaining immutable audit trails, separating concerns between metering, entitlements, and billing, ensuring deterministic enforcement, facilitating point-in-time reconstructions, and controlling change management to meet compliance requirements.