Zapier User Deploys AI Agent to Research Candidates and Draft Personalized Recruiting DMs.

Alex Lieberman, founder of enterprise AI firm Tenex, has built a Zapier agent that researches engineer candidates and drafts personalized recruiting messages based on their actual work.

The workflow, detailed in a LinkedIn post this week, triggers when Lieberman adds a LinkedIn profile URL to a Google Sheet. The agent then conducts web research to find specific, credible hooks, such as open-source projects, technical talks, or shipped products, and uses that context to generate a sub-100-word direct message in Lieberman's voice. Drafts are saved to a separate Google Sheet and posted to Slack for review before sending.

The agent includes a manual review flag. If it cannot identify a specific reason to reach out based on the candidate's public work, the workflow marks the row for human intervention rather than generating a generic message. Lieberman reported that of the first 10 drafts, six were sent as written, two required minor edits, and two were correctly flagged for manual review.

The setup reflects a growing pattern of agentic workflows in recruiting and sales, where automation handles research and drafting but keeps humans in the loop for final approval. Lieberman noted that top AI engineers receive dozens of recruiter messages weekly, making personalized, well-researched outreach critical to cutting through noise.

The agent's instructions specify a multi-step process:

• Trigger on new row in "FDE - Tenex" Google Sheet containing a LinkedIn URL
• Perform web search to gather data on current role, past employers, technical stack, open-source contributions, talks, and writing
• Identify the single most specific reason the candidate would be a fit, such as a particular project or client-facing technical work
• Draft a LinkedIn message under 100 words that leads with Lieberman's credibility (Morning Brew background, Tenex role), references the specific work, and ends with a soft ask
• Save draft to "FDE Outreach Drafts" Google Sheet with columns for name, LinkedIn URL, hook, and message
• Post notification to #recruiting Slack channel with candidate name and link to draft

The workflow prioritizes specificity over volume. Lieberman emphasized that he would rather skip a candidate than send a message without a credible angle.

Zapier has been expanding its agent functionality over the past year, including releasing open-source GTM agent skills built from real team use cases. The platform's agent builder allows users to define triggers, actions, and conditional logic in natural language, then generates a multi-step recipe that can be tested and refined.

Lieberman made the agent recipe available for others to copy and adapt. The post was tagged as a Zapier partnership.