# Vercel COO Reveals Playbook for Scaling AI Agents in Enterprise Vercel replaces manual workflows with autonomous agents, slashing lead qualification teams and support costs. - **Source:** [SaaStr](https://www.saastr.com) - **Published:** 2026-11-06 - **Topic:** [technology](https://brifing.si/c/technology) - **Original article:** - **On Brifing:** https://brifing.si/s/7946277b-6d06-481c-b667-f278d0b8dde6 ## Scaling Through Autonomous Agents Vercel has transitioned from manual processes to AI-driven operations under COO Jeanne DeWitt Grosser. By deploying internal agents, the company reduced its lead qualification team from ten people to one. This shift demonstrates how specialized agents can handle complex, high-volume tasks while maintaining performance levels comparable to top-tier human representatives. ## Economic Impact of Automation The lead qualification agent operates for under $5,000 annually in infrastructure and token costs. This represents a 32x return on investment compared to the previous ten-person team. Vercel achieved this by dedicating only 20% of one engineer's time to maintain the system, proving that high-impact automation is accessible for lean teams. ## The Three-Person Tripod Model Vercel builds every internal agent using a consistent three-person team: a GTM engineer, a data scientist, and the best performer in the target function. They document the expert's workflow, turn it into repeatable steps, and refine the model through shadow mode. This collaborative approach ensures agents mirror high-quality human decision-making processes. ## Technical Support Agent Performance Vercel’s internal support agent now manages 93% of total case volume, including deeply technical inquiries. Built by three engineers in two months, the system costs approximately $150,000 per year. This in-house solution outperforms generic vendor alternatives, allowing Vercel to maintain control over its customer service architecture while significantly reducing operational overhead costs. ## API-First Architecture Requirement Vercel mandates that all software tools must have accessible APIs to remain in their stack. Because agents interact via code rather than user interfaces, tools lacking developer surfaces are replaced. This headless design allows agents to perform complex tasks like updating CRMs and drafting outreach without requiring human intervention in the UI. ## Data Foundation for Agents The company’s most utilized agent, d0, allows employees to query business data in plain English via Slack. It operates on a semantic layer that translates questions into SQL, delivering answers in under a minute. This infrastructure prevents model hallucinations and ensures that all automated actions are grounded in accurate, real-time business metrics. ## Scaling Infrastructure Challenges Vercel notes that costs can jump 100x to 1,000x when moving from pilot to production volume. Traditional cloud infrastructure, designed for simple request-response cycles, often struggles with long-running agent jobs. Vercel utilizes its own Fluid compute technology to optimize these processes, cutting early adopter compute costs by up to 85%. ## Strategic Implications for Vendors The rise of internal agents forces B2B vendors to reconsider their value propositions. If a tool cannot be integrated into an automated workflow, it faces the risk of being cut by a customer's internal agent. Vendors must now provide robust developer platforms to remain relevant in an increasingly automated, agent-driven enterprise software landscape. ## Sources covering this story - https://www.saastr.com/top-5-learnings-from-vercels-coo-10-sdrs-down-to-1-a-5k-a-year-lead-agent-and-93-of-support-cases-handled-by-ai/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._