## Why Pelica exists

A value-based care team of eight can be accountable for a hundred thousand members. The data they need is scattered across 8 to 15 vendor portals, SFTP drops, and spreadsheets. The analytics layer is mature: most teams already know which gaps are open, which suspects are unconfirmed, and which members are slipping on adherence. The bottleneck is not knowing. It is doing, at volume, before the submission window closes.

Pelica was built to close that gap. Instead of one more dashboard, we put an AI workforce next to the operators: it makes the outreach call, books the visit, drafts the point-of-care documentation prompt, reads the MAO-004 rejections, and operates payer portals the way a staff member would. Every action carries its evidence, its provenance, and an audit trail, because in this market an undocumented action is a liability.

## The people building it

Pelica is built by engineers and clinicians who shipped AI at scale, guided by the operators who run risk-bearing contracts every day.

Catherine Zhao  
Co-Founder & CEO  
Former Google product lead for Responsible AI, Enterprise AI Agents, and the Imagen and Veo generative models. Selected for Google's APM program. Clinical AI researcher at Stanford School of Medicine and a medical assistant at the SF Free Clinic. Dartmouth.

Lalit Kundu  
Co-Founder & CTO  
Former Staff Engineer and Engineering Manager at Google and YouTube, area tech lead for a 45-plus engineer team in YouTube Commerce Billing. Forbes Technology Council member. Wharton MBA. ACM-ICPC World Finalist, the top 0.1% of competitive programmers.

## What we believe

01  
### Execution beats insight

Dashboards describe the problem. Value is created when the work gets done. We build the layer that closes the loop, not another report that asks a person to.

02  
### One record, every team

Risk, quality, pharmacy, network, and care management see the same member. One canonical record removes the coordination tax and the duplicate outreach that comes with 8 to 15 point vendors.

03  
### Provenance by default

Every recommendation is documentation-checked, RADV-defensible, and submission-window aware. An action you cannot defend in an audit is not a feature, it is exposure.

## By the numbers

Figures from live deployments, including our flagship physician-led IPA in New York.

175,000+  
Patients managed live on Pelica  
~10 hrs  
Saved per user per week at our flagship IPA  
96%  
Adherence on the three triple-weighted Part D measures  
2 to 4 wks  
From kickoff to a live copilot, not a multi-quarter project

## Backed by

Pelica is a Y Combinator company, supported by investors who back technical founders in healthcare.

Y Combinator  
Surge  
Point Capital  
Upside Partnership  
ISEED Ventures  
Progressive Ventures

## Operator-grade notes on value-based care.

V28 traps, Stars cut points, RADV patterns, and what we ship. A couple of emails a month, no fluff.