The Agentic Work Unit

2026-03-02 · Source: The Business Engineer · Field: Business & Management — Corporate Strategy & Leadership, Project & Product Management · Depth: Advanced, quick

Summary

Salesforce's Q4 FY26 earnings report highlights a significant shift in the SaaS industry's value denomination, moving beyond the traditional per-seat human user model. The company processed 19 trillion tokens, delivered 2.4 billion Agentic Work Units (AWUs), and managed 29,000 agent deals, alongside its existing seat-based business serving over 150,000 customers. This indicates a new equilibrium where SKU upgrades (seats) and consumption credits (work units) are split 50/50. The AWU represents a major software vendor's initial attempt to define a new unit of value when software users are AI agents rather than humans, prompting the entire industry to reconsider its economic structure.

Key takeaway

For SaaS entrepreneurs and investors evaluating future revenue models, recognize that the "per-seat" pricing paradigm is evolving. Your revenue forecasting and product strategy should account for agentic work units (AWUs) as a primary driver of value, potentially leading to a 50/50 split between traditional seats and consumption credits. Begin exploring how your offerings can be priced and consumed by AI agents to adapt to this structural economic shift.

Key insights

The SaaS industry's economic model is shifting from human-centric per-seat pricing to agentic work unit consumption.

Principles

• Software value is increasingly tied to agent output.
• The per-seat model is being challenged by AI agents.

In practice

• Evaluate agent-based consumption models for SaaS.
• Track agentic work units as a new performance metric.

Topics

• Agentic Work Units
• SaaS Business Models
• AI Agent Economics
• Software Pricing Models
• Salesforce Earnings

Best for: Investor, Entrepreneur, VP of Engineering/Data, Director of AI/ML, AI Product Manager, CTO

Related on AIssential

• The SaaS Destruction Map — The SaaS market is repricing due to AI agents breaking the per-seat pricing model.
• Salesforce & The Agentic Cannibalization — Salesforce is strategically self-cannibalizing its SaaS model to thrive in an agentic AI-driven economy.
• London's Airspeed raises €17.2 million Series A to build AI-powered execution layer for revenue teams - EU-Startups — Airspeed's AI agents provide a "system of action" for revenue teams, closing the loop between data and execution.
• PagerDuty CIO on the shift from seat-based to usage-based pricing for AI — AI integration is driving a shift from seat-based to usage-based SaaS pricing for better value alignment and cost reflection.
• The SaaSpocalypse Continues — AI disruption is fundamentally altering software market valuations and business models, shifting focus from seat-based subscriptions to agent-driven services.
• Intelligence Per Dollar — AI model evaluation is shifting from raw performance to cost-efficiency, prioritizing "intelligence per dollar."

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Business Engineer.