Company profile · 3 min read
Perfectly: the AI recruiting OS
Perfectly automates sourcing, outreach, screening and recruiting coordination for startups with urgent hiring needs.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
Perfectly is an AI-native recruiting agency for startups that need a hiring pipeline, not another pile of resumes. It fits high-growth teams with urgent roles and a recruiter or hiring manager who is willing to calibrate the system and review candidates.
What it does
The product automates sourcing, outreach, screening and qualification. The company profile says its recruiting agent, Paul, works in Slack and learns from candidate feedback. Its public material claims 4× faster hiring, 10× candidate volume, 2× higher interview pass rates and a reduction in recruiting work from 20 hours to one. Those are vendor claims. The homepage also presents a customer story in which one startup says it dropped other agencies after seeing more interview volume; that is a vendor-selected testimonial, not an independent review.
Why I’d look closer
The founders bring a clear technical thesis. The YC profile describes Victor Luo as a former ML scientist at TikTok and a repeat founder, Zhuang (Gary) Luo as ex-Meta and TikTok ML engineering, and Huimin Xie as a former tech lead and senior ML engineer at TikTok. The company is applying recommendation-system thinking to candidate matching, which is a more coherent wedge than simply putting a chatbot on top of an ATS.
What I’d ask
The public launch offer is also clear: Perfectly says it will undercut a current agency by 50% for a first role, or offer a discount. Beyond that, public pricing is not exposed. The bounded search surfaced a YC launch post and founder hiring posts, not an independent review. A buyer should ask how candidate quality is measured, whether outreach is customized enough to protect the employer brand, how consent and privacy work, and what “interview-ready” actually means.
My editorial take
Perfectly is worth a controlled test for a startup with one painful role and a high cost of delay. It is not a replacement for judgment. Start with a role where the success criteria are observable, compare the shortlist against your current agency or internal process, and treat the company’s speed and pass-rate numbers as hypotheses until your own funnel moves.
Quick facts
| Field | Sourced detail |
|---|---|
| Accelerator | Y Combinator, Winter 2026 |
| Industry | B2B; Recruiting and Talent |
| Public website | perfectly.so |
| Product | AI sourcing, outreach, screening and recruiting coordination |
| Company-reported results | 4× faster hiring, 10× candidate volume and 2× interview pass rate |
| Commercial signal | Launch offer says 50% under a current agency for a first role; normal pricing is not public |
Sources checked
| Volatile field | Source | Checked |
|---|---|---|
| Membership/founders/results | YC profile | 2026-09-19 |
| Public company search | YC launch post on LinkedIn | 2026-09-19 |
Cohort context
Perfectly is listed in Winter 2026. In our 2026-09-18 directory snapshot, 126 of 199 listed companies in that cohort have YC’s primary industry label B2B (63.3%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:20:23.499Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Not observed in this response |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗