We scan new podcasts and send you the top 5 insights daily.
Bratton says customers worry that if external-agent fees are too high, they will skip Salesforce's connectors and run their own AI on top of Salesforce data. That would leave Salesforce more dependent on Agentforce, whose revenue is far below Claude Code and Codex usage, according to her reporting.
AI agents generate orders of magnitude more data than humans. The high storage costs of traditional CRMs like Salesforce, which can be 1000x more expensive than modern databases like Postgres, create a strong financial incentive for companies to move data out, weakening the CRM's moat.
To combat seat contraction, legacy vendors are charging extra for agent and API access. This short-sighted strategy risks alienating new customers who will refuse to adopt platforms that are not inherently 'agent-friendly' and limit automation, creating openings for modern competitors.
To offset declining seat-based revenue, some SaaS vendors are drastically increasing API prices. This strategy backfires with AI agents, which generate massive data volumes. The high costs create a powerful incentive for customers to migrate their data elsewhere, accelerating the vendor's own decline.
Enterprise platforms like Salesforce and Workday are constructing API toll gates to monetize third-party AI agents interacting with their systems. However, vendors face significant buyer pushback and have introduced grace periods before charging fees. Software providers must demonstrate clear added value beyond raw API access, because enterprise buyers strongly resist paying tolls simply to allow autonomous agents to interface with their own proprietary corporate data.
Instead of a full "rip and replace," large companies like Sanofi are keeping systems like Salesforce as a "system of record" but are moving significant workloads (up to 80%) to custom AI agent-driven processes. This subtly undermines the value and pricing power of incumbent SaaS vendors.
Laura Bratton reports Salesforce will make customers register outside agents, such as Claude or OpenAI, and pay when those agents call into Salesforce, even through customer-built MCP connectors. Pricing is due around November, and an unlimited headless license is said to be available, according to a consultant.
SaaStr's experience shows that while human user seats for Salesforce decreased dramatically, intensive data usage from 20 AI agents led to a significant net increase in their bill. This suggests a shift from per-seat to consumption-based pricing models driven by agentic AI.
The ability for customers to build their own software features using AI agents directly threatens the traditional SaaS upsell model. During negotiations, customers can now credibly threaten to "roll their own" features instead of paying for higher-priced tiers, weakening the vendor's pricing power.
Bratton calls Salesforce's move a turning point in the SaaSpocalypse debate: can software firms charge when customers use their own AI to reach data they already pay to store? Customers are waiting for pricing before committing, because the answer decides whether toll gates work for many SaaS companies.
While Salesforce seems difficult to disrupt externally, its large Fortune 500 customers have the resources to build their own tailored solutions using AI. They can bypass paying for a bloated software suite they only partially use, posing a significant "insourcing" risk.