- Revenue Operators
- Posts
- Signal #04: The AI Sales Agent Pricing War Begins
Signal #04: The AI Sales Agent Pricing War Begins
Strategic intelligence on the operator vs theater divide
Olivier Madel
October 07, 2025
Hey, Revenue Operators,
The AI sales agent pricing war just went live. Salesforce's Agentforce launched October 1st with three pricing models, Meta dropped Business AI for SMBs on October 2nd, and acquisition activity shows enterprises buying capabilities they can't build fast enough.
But here's the pattern most people miss: Half of GTM teams use AI weekly, yet only 21% have redesigned workflows. That gap isn't a bug… it's the exact arbitrage opportunity creating 30% salary premiums for operators who build systems versus reps who buy tools.
Today's signals reveal which capabilities enterprises are desperate to acquire, the exact workflow redesign skills commanding premiums, and your positioning move while everyone else chases shiny AI objects.
Signal 1: Salesforce's Triple-Threat Pricing Model
October 1, 2025
illustration Midjourney
Salesforce launched Agentforce with three distinct pricing models: $0.10 per agent action via flex credits, per-conversation pricing for dialogue-based workflows, and $125-550/user/month for unlimited usage tiers. This fragmentation signals market uncertainty about AI agent unit economics. Early operator analysis shows the per-action model favoring high-volume/low-complexity tasks, while unlimited tiers make sense for strategic conversation design workflows. The pricing structure reveals Salesforce betting on both commoditized automation and premium human-AI orchestration.
What this means: The market hasn't settled on AI agent value metrics. Companies experimenting with multiple pricing models creates opportunity for operators who can calculate true TCO across different workflow types and prove ROI using actual usage data rather than vendor projections.
Operator Move: Map 3-5 of your current manual workflows. Calculate potential action volume for each if automated. Model costs across per-action, per-conversation, and unlimited pricing to identify which model fits your actual usage patterns. Build this pricing evaluation skill now… it's becoming core operator capability.