SaaStr runs its sales and operational functions with roughly three human employees supported by around 20 AI agents, according to a recent post by Brad Blumberg on LinkedIn. The company previously used up to 30 agents but consolidated due to overlapping functions and conflicting outputs. These AI agents perform real operational roles and are integrated with actual systems, replacing much of the manual work in the sales organization.
The AI agents at SaaStr evolved from dashboards, project management tools, and websites into autonomous agents as the company increasingly relied on them. About six of these agents are used daily, handling tasks such as daily revenue tracking, forecasting, and campaign performance monitoring. The consolidation was necessary because overlapping agents produced conflicting answers, making reconciliation difficult. The company has also documented what each agent refuses to do and where they have failed, providing transparency on their limitations.
This approach highlights the growing trend of integrating AI agents into SaaS operations to improve efficiency and reduce human workload. SaaStr’s experience reflects challenges in managing multiple AI tools, including agent sprawl and conflicting outputs, which parallels issues seen in SaaS product sprawl. The company’s model of combining a small human team with AI agents could serve as a reference for other SaaS firms aiming to scale operations with limited staff.
Brad Blumberg’s LinkedIn post detailing SaaStr’s AI agent usage was published this week, providing a comprehensive overview of the company’s current operational setup and the evolution of its AI tools.