G2’s 2026 AI Agents Insight Report surveyed 1,019 B2B decision-makers in July 2026 and found that only 9% of buyers consider per-seat pricing the fairest model for AI agent software. Instead, 57% prefer pay-as-you-go models based on consumption or completed tasks, marking a significant shift from 2025. Additionally, 74% of buyers are willing to consider switching vendors at renewal for stronger AI agent capabilities, highlighting changing buyer expectations in AI software pricing, according to saastr.com.
The report reveals that 86% of B2B software buyers purchased AI agent software or services in the past year, up from 69% in 2025, with 66% having agents in limited or scaled production. Budgets for AI agents now exceed $1 million annually at about half of companies surveyed. While many companies initially added AI agents under existing per-seat plans or added per-seat AI SKUs in 2025, this approach is losing favor as buyers increasingly demand pricing models aligned with actual agent usage and outcomes, saastr.com detailed.
This shift matters as AI agents operate continuously without human logins, making per-seat pricing less relevant. The rise in pay-as-you-go and pay-per-outcome models reflects buyers’ desire for pricing that matches the value delivered. The report also notes a decline in fixed pricing preference, from 35% in 2025 to 21% in 2026. These trends suggest that vendors relying on traditional per-seat pricing risk losing business, with 30% of buyers ready to switch vendors for better AI capabilities and 44% open to piloting alternatives, saastr.com reported.
The report’s findings underscore the urgency for SaaS vendors to reconsider AI agent pricing strategies before 2027 renewals. With agent budgets surpassing $1 million at many companies and 74% of buyers open to switching vendors, pricing models that reflect actual usage and outcomes will be critical. G2’s 2026 AI Agents Insight Report was based on responses from 1,019 B2B decision-makers surveyed in July 2026, saastr.com confirmed.