We asked an agent to log 40 messages on 3 CRMs via MCP. One cost 4x as much (goodtiming.ai)

In a recent test, an AI agent logged 40 messages across three CRM platforms—Attio, HubSpot, and Atomic CRM—revealing significant differences in efficiency and cost. Attio's CRM took 240 seconds and 117 API calls for the task, resulting in a total cost of $1.86, while HubSpot streamlined the process to 177 seconds and 21 calls for a cost of $1.42. In stark contrast, Atomic CRM completed the same task in just 83 seconds with only 10 calls, costing a mere $0.49. This comparison highlights the effectiveness of API design and the importance of minimizing redundant requests in reducing operational costs. The findings underscore a critical concern for the AI/ML community: the optimization of API performance can lead to substantial cost savings and improved user experience. As organizations rely on CRM systems for managing interactions, the efficiency of these systems can directly impact resource allocation, especially for companies with extensive outreach operations. By using a framework called Baton to analyze the calls made during the logging process, it was evident that while Attio's approach created excessive API traffic, Atomic CRM's design allowed for batch processing of requests. This emphasizes the necessity for CRM providers to enhance their API capabilities, making them more efficient for AI agents, ultimately leading to reduced overhead costs in high-volume environments.