By Jacques Bughin and Philipp Remy
As autonomous software performs more operational work, enterprise pricing is shifting from user access toward measurable execution and outcomes.
Enterprise software is entering a structural transition as agentic artificial intelligence changes how value is created and monetized across digital systems. In this article, Jacques Bughin and Philipp Remy argue that traditional seat-based SaaS pricing no longer reflects how software increasingly operates inside organizations. As autonomous agents begin executing workflows, processing decisions, and delivering measurable operational outcomes, software companies face growing pressure to redesign pricing around execution itself. The article explains why this shift matters strategically and how SaaS firms can move toward new monetization architectures without losing control of value creation.
For more than two decades, enterprise software has been built around a deceptively simple premise: software amplifies human productivity. The Software-as-a-Service (SaaS) revolution monetized this premise through seat-based subscriptions. Firms paid for access to digital tools used by employees, and vendors scaled revenue as organizations hired more people.
That economic architecture is now breaking down. The emergence of agentic artificial intelligence marks a structural shift in how software creates value. Increasingly, software no longer assists human workers—it performs the work itself. Autonomous agents can interpret requests, execute workflows, trigger transactions, and learn from operational feedback loops. In doing so, they transform software from an interface layer used by people into an execution layer embedded directly in operations.
This transition signals the emergence of a new economic paradigm: the execution economy.In this new architecture, value is no longer measured by the number of users interacting with software interfaces. Instead, value is determined by the number of operational decisions executed autonomously, the tasks completed, and the outcomes delivered. The traditional monetization anchor of the SaaS era—the human seat—begins to dissolve.As artificial agency expands, software pricing must evolve accordingly. Vendors are moving from access-based pricing to architectures that monetize execution, performance, and delegated authority.The shift represents the most significant transformation in the software industry since the rise of cloud computing.
Why the Fastest-Growing Software Businesses are no Longer Saas
The most successful software companies of the last decade share a surprising characteristic: they are not monetized like traditional software at all. Stripe charges businesses only when a payment is successfully processed. AWS Lambda charges only when a piece of code executes. Snowflake charges only when data is queried or processed. Anthropic’s Claude charges when artificial intelligence performs reasoning tasks. None of these companies primarily sell seat licenses. Instead, they monetize execution. Stripe earns revenue from completed financial transactions. AWS Lambda bills for the execution of functions. Snowflake charges for computational queries and storage consumption. Claude generates revenue from tokens processed during reasoning and task execution.
In each case, the customer does not pay for access to software. The customer pays when the software performs work. This distinction may seem subtle, but it represents a fundamental shift in the economic architecture of enterprise software. For over two decades the software industry has been organized around what might be called the interface economy. Software was delivered through user interfaces and monetized through subscriptions tied to human access. The Software-as-a-Service model became the dominant paradigm because it aligned pricing with the number of employees using a system.
The Missing Layer: Downstream Execution Software
The four companies mentioned above illustrate the economic logic of this new model. Stripe functions as a transaction utility. Its revenue grows with the economic throughput of the businesses using it. As merchants sell more products, Stripe processes more payments and captures a percentage of each transaction. AWS Lambda operates as a compute execution utility. Developers deploy functions that execute in response to events, and AWS charges for the duration and memory consumed by each execution. Snowflake monetizes data computation. Companies pay when data is processed, queried, or transformed rather than when users log into a software interface. Anthropic’s Claude monetizes artificial intelligence execution. Customers pay for tokens processed and reasoning cycles performed by the model. Although these companies operate in different domains—payments, cloud infrastructure, data platforms, and artificial intelligence—they share the same economic architecture.
Revenue scales with work performed by the system, not with software seats.
Traditional SaaS companies monetize human productivity. A firm with one hundred employees using a system typically purchases one hundred licenses. Revenue therefore scales with workforce size. Execution utilities operate under a completely different economic model. Revenue scales with operational throughput .If a company processes more payments through Stripe, runs more compute tasks through AWS, executes more queries in Snowflake, or performs more reasoning cycles using Claude, the platform’s revenue increases automatically. This creates a powerful economic dynamic. In the SaaS model, increasing software efficiency often reduces revenue because fewer human users are required to perform the same work. In the execution model, increasing software efficiency increases revenue because the system performs more operational tasks. The platform monetizes activity and outcomes, not access.
However, most enterprise software companies are not positioned like Stripe, AWS, or Snowflake. These firms operate at the utility layer of the technology stack. They provide foundational infrastructure that executes tasks but does not control the business workflows in which those tasks occur.
The next wave of transformation will occur downstream, where traditional SaaS companies operate. And the option value can be rather high: When software performs the work itself, revenue scales with the volume of work performed, not with the size of the workforce.A customer support platform that resolves ten million cases per year can generate significantly more revenue than a traditional ticketing system priced per agent seat. Similarly, a financial automation platform that processes billions of invoices may capture a share of operational spending far larger than the subscription fees charged by legacy accounting software. In other words, AI systems expand the economic boundary of the software industry. Instead of selling tools used by workers, software vendors increasingly sell execution capacity that replaces workers.
The Strategic Question for SaaS
This raises the defining strategic question for the next generation of enterprise software: How can traditional SaaS companies transition from selling software interfaces to monetizing execution outcomes?
Despite this massive opportunity, most existing SaaS companies are not yet positioned to capture it.Many products remain interface-centric systems designed to assist human users rather than execute operational workflows. These companies face a strategic dilemma. If they retain seat-based pricing, they risk revenue deflation as automation reduces the number of human users.If they attempt to adopt outcome-based pricing without controlling execution, they assume risks they cannot manage.The transition therefore requires a deeper transformation. SaaS products must evolve into execution platforms capable of performing measurable operational tasks autonomously. Only then can vendors monetize outcomes rather than access.
Once this authority exists, the next challenge is strategic defensibility. In a world where general-purpose AI agents such as Claude or open-source frameworks like OpenClaw can perform tasks autonomously, execution capabilities risk becoming commoditized. Only after these two conditions are satisfied—operational authority and defensibility—can firms safely transition toward outcome-based pricing models that capture the economic value of digital labor.
The journey therefore unfolds in three stages.
Stage 1: Acquiring orchestration and execution power.
Traditional SaaS companies typically operate at the interface layer of enterprise software. They provide dashboards, analytics, or workflow tools that assist human decision-makers but rarely control the underlying operational processes. Outcome-based monetization is impossible in such environments because the software does not control the outcome.The first step toward the execution economy is therefore acquiring orchestration power. Orchestration power exists when a system coordinates multiple enterprise services and determines the sequence of actions required to complete a workflow. Execution power exists when the system can perform those actions autonomously. Companies can acquire these capabilities through three architectural moves.
- Workflow Ownership. The most important source of execution authority is ownership of the workflow itself. Platforms that sit at the center of operational processes gain the ability to coordinate multiple enterprise systems simultaneously. For example, an HR platform that manages employee onboarding orchestrates identity management, payroll setup, access provisioning, and compliance verification. Once the platform orchestrates these interactions, it can gradually automate them.
- System Integration Depth Execution power increases with integration depth. Software that merely exports analytics cannot perform operational actions. By contrast, systems with direct API access to enterprise infrastructure—payments, authentication systems, inventory databases, or logistics platforms—can execute decisions automatically. Integration therefore transforms software from an advisory system into an operational system.
- Guardrailed automation Execution authority should initially be introduced through guardrail automation. In early stages, systems propose actions that require human approval before execution. Over time, as the system demonstrates reliability, approval thresholds can be relaxed and workflows can become fully autonomous. Guardrailed automation allows vendors to gather performance data necessary to justify outcome-based pricing. The critical insight is that execution authority is rarely obtained through a single technological breakthrough. It emerges gradually as software becomes embedded within operational workflows.
Stage 2: Defending execution power against general AI agents
Once software acquires execution authority, the next strategic challenge is preventing that authority from being captured by general-purpose AI agents. Large language models and open-source agent frameworks can often replicate generic task execution capabilities. If enterprise workflows are easily accessible to these agents, the vendor that owns the software interface may lose control of value creation. Defensibility therefore depends on owning assets that general-purpose agents cannot easily replicate.
Two sources of defensibility are particularly important.
- Data Context and Operational Memory Execution systems accumulate large volumes of operational data that describe how workflows behave in real-world environments. This data includes historical transactions, process exceptions, policy constraints, and organizational preferences. Over time, this operational memory becomes a powerful asset because it allows the system to optimize decisions in ways that generic AI agents cannot replicate without access to the same data.
- Embedded Governance and Compliance Enterprise workflows often operate within strict regulatory and governance frameworks.Execution platforms that embed compliance rules, audit trails, and policy enforcement mechanisms become deeply integrated with organizational risk management processes. These governance layers create switching costs because replacing the platform would require rebuilding regulatory infrastructure.
Stage 3: Monetizing Outcomes
Once a platform controls execution authority and possesses defensible assets, it can begin transitioning toward outcome-based monetization. The key principle of outcome-based pricing is that customers pay only when the platform produces measurable economic value. However, implementing such pricing requires a carefully designed architecture that balances customer trust with vendor upside.
Three mechanisms are critical:
- Proof of outcome. Customers must be able to verify that the platform has successfully performed the task being billed. Execution systems therefore generate verifiable records of each decision and action performed. These records provide an auditable trail demonstrating that outcomes were achieved as claimed. Proof-of-outcome mechanisms transform autonomous systems from opaque algorithms into accountable operational infrastructure.
- Risk-sharing contracts. Outcome-based pricing typically involves some form of risk sharing between vendor and customer. In many implementations, vendors charge only for successful outcomes while absorbing the computational cost of failed attempts. This arrangement aligns incentives and encourages customers to adopt automation without fear of paying for unsuccessful results.
- Expansion through operational throughput. Once trust has been established, outcome-based pricing unlocks significant expansion opportunities. Unlike seat-based pricing, which is limited by workforce size, outcome-based pricing scales with the volume of operational work performed by the platform. As customers expand their business operations—processing more transactions, serving more customers, or executing more workflows—the platform’s revenue grows automatically. In effect, the vendor captures a share of the digital labor market created by autonomous systems.
Conclusion
The transition from the interface economy to the execution economy represents a fundamental redefinition of software’s role in the enterprise. For decades, software monetization relied on access: the number of users interacting with digital tools. In the agentic era, value migrates toward execution—the ability of autonomous systems to perform operational decisions continuously and reliably. This transformation reshapes pricing models, governance frameworks, and competitive dynamics across the software industry. Companies that remain tied to seat-based economics risk falling into the deflationary trap of automation. Those that successfully transform software into an execution platform will capture a share of the operational flows that define modern digital economies.
In the age of artificial agency, the most valuable software will not be the software that workers use. It will be the software that works, and that is priced accordingly for the value it brings.
Jacques Bughin
Philipp Remy
