AI agents killed Marketo's pricing model: cut per-call costs or lose usage

SaaStr ran 20+ AI agents, hit a vendor 45,000 times daily, then cut usage 90% because the bill was too high. The lesson for B2B SaaS: agents measure every API call. If your product costs too much per action, they route around you. Same day.

AI agents killed Marketo's pricing model: cut per-call costs or lose usage

The Problem

AI agents do not buy software and forget about it. They track cost per API call, cost per task, cost per outcome. Every day.

SaaStr runs 3 humans and 20+ AI agents in production. One core vendor saw their bill jump 40% this year, almost all from AI usage. The agents were hitting that system 35,000 to 45,000 times daily.

SaaStr cut those calls by 90%. The need did not change. The price made the agent route around the vendor.

Marketo is gone. Most marketing email now runs through a cheaper API that an AI agent set up and migrated to. SaaStr is also moving off some of its core AI products. They work fine. They cost too much per unit of usage for agents that need to call them thousands of times a day.

The Math

A human logs in a few times a week. An agent calls you thousands of times a day, with no seat limit and no vacation.

Cut your price per action by 80%, get 20x the agent usage, and revenue from that customer goes up 4x. You also end up inside a workflow that runs all day instead of a tool someone opens on Tuesdays.

Another 8% uplift at renewal will not get you there. Lower unit prices through higher usage volume will.

The Databricks Playbook

Databricks charged $1,000 per user on top of consumption early on. Customers told CRO Ron Gabrisko that the per-user fee was why they were not rolling Databricks out more widely. He made users free. Consumption went way up.

Databricks is now a $200B company. The per-user fee was holding back far more revenue than it brought in.

With agents, this matters more. A $1,000 seat fee slowed down humans. Per-call pricing that is too high throttles agents every single day.

What This Means for Sales Teams

If you are selling into accounts running AI agents, expect procurement to track API costs daily, not annually. Expect them to optimise around your pricing if it is too high. Expect usage to drop mid-contract if the per-call economics do not work.

If your company is refounding around AI: HubSpot announced a 7% workforce reduction in October 2024, citing its shift toward AI-driven customer outcomes and changing product, pricing and service models. Co-founder Brian Halligan returned as CEO in November. The company grew from $1.3B in 2021 to an expected $3.7B in 2026 while expanding into agentic workflows and Breeze, its AI layer.

Refounding is not just product. It is pricing, operating model, quota structure and what your AEs are actually selling. Agents do not care about seat counts. They care about cost per outcome.

The vendors that win in agentic B2B will grow contract value through usage volume at lower unit prices. The ones that do not will watch agents route around them, one API call at a time.