The rise of AI-powered agents managing CRM interactions has created unprecedented data volume spikes in Salesforce, highlighted by one operator writing 40GB of agent-generated data in just 30 days. This shift not only strains traditional platforms’ pricing models but also reshapes partnership dynamics, as demonstrated by ServiceTitan’s recent disconnection of Podium after agents started displacing core system functions.
- Agents generated 40GB of data in Salesforce within a month, mostly via API.
- ServiceTitan ended its Podium partnership as agents began replacing core system functions.
- Agent-driven CRM usage raises major costs and calls for new data handling approaches.
What happened
In the latest episode of The Agents, a SaaS AI operator revealed that their deployed AI agents had generated about 40GB of data inside Salesforce in just 30 days without anyone manually entering it. These agents continuously log tasks, emails, call data, and enrichment updates through Salesforce’s API, causing data volumes—and subsequently storage costs—to escalate dramatically. The scale equates to roughly 21 million individual records, an order of magnitude higher than traditional human user activity.
Meanwhile, ServiceTitan, a major system of record provider for trades, terminated its integration with Podium, a partner turned competitor. Podium used AI agents to automate customer interactions, appointment scheduling, and job tracking traditionally managed by ServiceTitan. This agentic shift blurred partnership lines, prompting ServiceTitan to sever ties after nearly a decade, citing displacement concerns through agent-driven systems.
Why it matters
This development highlights that legacy CRM systems and pricing models were designed for infrequent human interactions, not continuous machine-generated data streams. Salesforce’s architecture struggles under the volume and frequency of AI agent writes, causing unexpected cost overruns and usage alerts. For SaaS operators, the implications are significant: AI agent usage can rapidly multiply data volumes leading to exponentially higher platform expenses and complex API cost dynamics.
Moreover, Podium’s experience exposes a new competitive risk embedded in agent deployments: partners empowered by agents risk evolving into direct competitors by capturing end-user workflows and data within system of record platforms. This forces system providers to reconsider their integrations and competitive strategies amid a rapidly agentified SaaS ecosystem.
What to watch next
Expect growing scrutiny on how AI agents impact system of record costs and partnerships in SaaS. Many operators will investigate or migrate to alternative data storage solutions like Postgres-based platforms, which can handle large data volumes more cost-effectively. Monitoring how Salesforce and others adapt pricing and architecture to agent-driven volumes will be critical to sustaining AI-enabled growth.
Keep an eye on emerging agent uses in customer experience (CX) and support, where agents embedded on websites or apps may bypass traditional CRM altogether. This could further disrupt traditional system of record roles and integration models. Meanwhile, enterprises will balance AI automation benefits with the need for scalable, outcome-based pricing models rather than metered API usage fees.