Zapier Pilots Internal AI Spend Tracking Tools as Compute Costs Accelerate.

Zapier is testing new internal tools to help its employees understand and manage rising AI costs, according to a LinkedIn post from Brandon Sammut, the company's director of product.

The automation platform recently piloted two internal skills with a handful of teams: /ai-usage-report, which provides a 30-day snapshot of individual AI tool usage and spending, and /ai-spend-coach, a guided walkthrough designed to identify quick optimization opportunities. Starting next month, every Zapier employee will receive a personalized monthly report.

The move reflects a growing tension companies face as AI adoption spreads. Sammut noted that employees wanted visibility into their AI spending before being asked to make decisions about which tasks warrant expensive frontier models versus cheaper alternatives. "You can't ask folks to fine-tune spend they can't see," he wrote.

Zapier's approach appears deliberately cautious. The company has encouraged employees to experiment with AI tools, but now faces the challenge of managing compute bills that Sammut described as growing "fast." He acknowledged the trade-offs: heavy-handed cost controls risk stifling innovation, while no controls at all mean missing opportunities to teach efficient AI usage.

The company is still determining optimal usage patterns. "Overall, it's pretty early to understand the optimal usage patterns," Sammut said, suggesting this pilot is part of a longer-term effort to figure out sustainable AI spending practices.

The /ai-usage-report skill surfaces spending data that was previously scattered across multiple platforms. By consolidating 30 days of usage into a single view, employees can see where their AI budget is going, which tools they're using most, and how their spending compares to colleagues.

The /ai-spend-coach takes a more prescriptive approach, walking users through their usage patterns to spot areas where they might switch to less expensive models or reduce unnecessary API calls. The guided format suggests Zapier is betting on education rather than top-down restrictions.

Both tools were developed based on employee feedback requesting centralized spending visibility. The pilot phase allowed Zapier to refine the reporting format and coaching suggestions before rolling them out company-wide.

Zapier's internal experimentation comes as the company has been building and releasing open-source AI agent skills based on real team workflows. The company also recently launched Zapier MCP, which lets Claude access thousands of apps without custom API integrations.

These product developments increase AI usage internally, making cost tracking more urgent. As companies like Zapier expand their AI capabilities, they face the same optimization challenges they're trying to solve for customers.

Zapier CTO Bryan Helmig will address AI cost management in a live Q&A on June 24, covering cost per task across models and how to determine which automation work is worth doing. The session suggests Zapier plans to share what it learns from this pilot more broadly, both internally and with its user base.