ServiceTitan ended its nine-year partnership with Podium, affecting about 1,000 shared customers, after Podium adopted an agentic system approach that significantly increased data usage in the system of record. Podium’s agents wrote roughly 40GB of data into Salesforce within 30 days, a jump from 5GB, triggering storage overages and raising questions about the safety and cost of AI-driven data management, according to saastr.com.
The shift to agentic systems means software agents continuously write data such as task records, email interactions, call metadata, and enrichment directly into the CRM via APIs, without human logins. Podium’s agents generated approximately 21 million records in a month, causing storage costs to escalate unexpectedly. ServiceTitan, a major system of record for trades, decided to sever ties with Podium after the latter’s agent revenue crossed nine figures, reflecting tensions over data control and system usage.
This case highlights the challenges companies face as AI agents increasingly manage customer relationship data autonomously. The surge in data volume and storage costs can strain traditional CRM platforms like Salesforce. The dispute between ServiceTitan and Podium underscores broader industry concerns about integrating AI agents into core business systems and the financial implications of such transitions, especially for B2B SaaS providers handling large-scale data.
Podium’s agentic system wrote 40GB of data into Salesforce in about 30 days, with 99% of it through the API, while neither company’s staff logged into Salesforce during that period, saastr.com reported. This data volume and the subsequent ServiceTitan decision mark a pivotal moment in how AI agents interact with systems of record in SaaS environments.