Vercel's lead agent took inbound qualification from a 10-person team to about one person
Vercel built an agent that researches and qualifies inbound sales leads. A GTM engineer shadowed the best SDR, copied her workflow, and ran the agent in shadow mode before going live.
- Organization
- Vercel
- Industry
- Developer Platform
- When
- 2025
- Evidence
- Press report
- Agents and tools
- Custom agent (Vercel AI SDK)
Reported outcome
Inbound SDR work went from about 10 people to 1 plus a part-time person; the agent costs about $5,000 a year to run (as reported).
- ~10 → ~1SDR team
- ~$5,000/yrRun cost
- ~20% of one engineerMaintenance
Caveats
Reported by a news outlet, usually based on company statements.
- Numbers come from a conference talk by Vercel's COO, reported by SaaStr.
- Team-size figures differ between sources.
What the agent does
When a lead comes in, the agent researches the company and person, scores fit, and drafts a reply or routes the lead. A human checks edge cases.
How they built it
A three-person team: a GTM engineer, a data scientist, and the best SDR as the subject-matter expert. The engineer watched every tab the SDR opened for several days. The agent then ran in shadow mode, with the SDR reviewing its output, before it took over.
What to copy
- Copy your best person's workflow, not a generic one.
- Run in shadow mode until the expert agrees with most decisions.
- Pick work with high repetition and low judgment. Vercel calls this the sweet spot.
Starter prompt
Written by us, not by Vercel.
New inbound lead: {name}, {email}, {company}, message: "{message}"
1. Research the company: size, industry, tech stack signals, recent news.
Cite a URL for each fact.
2. Score fit 1-5 against this ICP: {icp}. Explain in 2 sentences.
3. Route: 4-5 → draft a reply offering a call; 2-3 → send the
self-serve resources email; 1 → no action.
4. If anything looks like spam or a competitor, flag it for a human.
Sources
- Vercel took a 10-person SDR team down to 1 — SaaStr, 2025
- What we learned building agents at Vercel — Vercel, Nov 2025
Added Oct 16, 2026. Spotted an error? Send a correction.
Get new use cases weekly
One short email. Sources included.
Related use cases
Anthropic's growth marketer cut ad creation from 30 minutes to 30 seconds with Claude Code
Work that took about 30 minutes per ad now takes about 30 seconds; the plugin can make up to 100 ad variations at once.
Amazon says its Rufus shopping assistant is on track for $10B+ in extra yearly sales
Amazon expects over $10B in annualized incremental sales; shoppers who used Rufus were 60% more likely to buy.