Salesforce's Agentforce Strategy: Accelerating Q2 FY27 Results, Changing ARR Metrics, and the Three-Model Pricing Hedge

Salesforce's Q2 fiscal 2027 results (reported August 26, 2026) signaled a powerful reversal of the growth slowdown that plagued the company in early 2026. Constant-currency cRPO growth accelerated to 14% year-over-year (reaching $33.5 billion), and the company raised its full-year FY27 revenue guidance to $46.1B–$46.4B. The primary driver of this renewed momentum was the rapid enterprise adoption of Agentforce and the Data 360 platform.

Redefining the AI ARR Metric

Salesforce reported that Agentforce ARR exceeded $1.5 billion, representing a 240% year-over-year increase. Combined Agentforce + Data 360 ARR reached nearly $3.9 billion. However, Salesforce quietly adjusted the scope of this metric in Q2:

Scope Expansion: Starting in Q2 FY27, the "Agentforce ARR" metric was expanded to include Salesforce's broader AI offerings, Slackbot, and Headless 360.
RevOps Tracking: Analysts note that this change helps paint a highly favorable picture of AI growth, though it obscures pure-play Agentforce contract values.

The Three-Model Pricing Hedge

To overcome the severe procurement friction of consumption-based AI pricing, Salesforce is now running three different pricing models for Agentforce simultaneously. This multi-model approach is designed to let customers self-select based on their buying preferences:

Per-Conversation ($2 per conversation): Launched in October 2024, this model was theoretically elegant but fell apart in production. Enterprise buyers could not model their costs (e.g., when a single query triggered multiple backend processes), and it priced out SMBs. This led to tepid early paid adoption (only 3,000 paid out of 5,000 deals).
Flex Credits ($0.10 per action): Launched in May 2025 (100,000 credits for $500), this model offered granular action-based pricing. However, it remained unpredictable for standard corporate budgeting.
Per-User Licenses ($125+/month): Launched in late 2025 as the Agentic Enterprise License Agreement (AELA). This model wraps the "digital workforce" into standard seat licenses, giving CFOs a predictable, budget-friendly contract shape they already know how to approve.¹

By running all three models in parallel, Salesforce successfully expanded its paid Agentforce customer base. However, the company is actively battling seat cannibalization. As its own customer service agents are replaced by Agentforce (which resolved 84% of Salesforce's 380,000+ support interactions), its core per-seat model is under pressure. Salesforce sales engineers report a 10% reduction in seats across 90 enterprise accounts as clients find they need fewer human agents to do the same work.

An instance of Pure outcome-based pricing collapses under the weight of enterprise budget unpredictability. — Salesforce is introducing seat-based pricing for its digital workforce because enterprise buyers demand the predictability of standard contract shapes over volatile, consumption-based fees. ↩︎

Part of

This finding is an example of a pattern recurring across your work:

Standardized context protocols must replace stateless APIs to coordinate agents across enterprise boundaries.

Shows a major incumbent implementing MCP to let outside agentic ecosystems seamlessly access its underlying data footprint.

Per-seat licensing collapses when software eliminates the human headcounts it used to price.

It shows Salesforce acquiring pricing infrastructure to native-bill consumption and outcomes as AI agents reduce paid seat logs.

Standalone AI point solutions collapse without ownership of the integrated system of record.

Salesforce's acquisition of Fin illustrates how standalone agent point solutions are systematically absorbed into integrated platforms that hold the underlying customer systems of record.

Human payrolls must be liquidated to finance the extreme capital costs of silicon.

Salesforce executed targeted layoffs to shift internal payroll expenses away from older SaaS products and directly into its core agentic AI development.

Per-seat software licensing collapses the moment AI agents replace human operators.

Enterprise giants are acquiring successful, outcome-priced AI agent startups to pivot their own models away from collapsing per-seat subscription streams.

The path to agent scale runs exclusively through incumbent distribution pipelines.

To capture early agentic momentum, enterprise software giants are aggressively acquiring independent startups to integrate their technologies into dominant distribution channels.

Revision history

Update Salesforce's Agentforce strategy note to incorporate Q2 FY27 earnings, the cRPO acceleration, the redefined ARR metric, and the concurrent deployment of three pricing models.

5d ago· by the agent

Update Salesforce note with the full financial and technology details of its $3.6B acquisition of Fin (formerly Intercom) announced on June 15, 2026.

54d ago· by the agent

Update Salesforce note with completed m3ter acquisition on July 1, 2026 and details of the $3.6B Fin (Intercom) acquisition.

61d ago· by the agent

Update Salesforce note with details of the Fin and m3ter acquisitions, June 2026 layoffs, and Agentforce's Q1 FY27 ARR metrics.

68d ago· by the agent

Update Salesforce note with the massive June 2026 acquisitions of Fin (Intercom) for $3.6B and m3ter, highlighting the shift toward consumption-based billing and defensive consolidation.

75d ago· by the agent