← SELECTED WORK
GOVERNANCE TECH

Quorate

An AI document-intelligence platform that reads proxy statements and investor voting guidelines, then flags governance compliance risk, built for a US corporate-governance analytics startup.

Product preview

Representative interface, rebuilt with fictional data. The client's production UI is not shown.

The challenge

Every proxy season, a public company's governance team has to answer the same question for each of its largest institutional shareholders: does our proxy statement comply with your voting guidelines, item by item? Each investor publishes its own dense policy PDF, and cross-referencing them against a 100+ page proxy statement across dozens of policy topics is weeks of manual analyst work, repeated for every investor, every year, under filing-deadline pressure. Quorate needed that entire workflow automated, without sacrificing the traceability that governance professionals require to trust an answer.

Our approach

Three-service architecture, a Next.js app, a FastAPI core, and a Python AI worker, coupled only through AWS SQS so model work ships independently of API work

Multi-stage document pipeline: OCR, ballot-item extraction, question answering, guideline extraction, and compliance flagging, with every answer citing source-PDF pages

Four LLM providers (Gemini, Claude, GPT, Mistral OCR) behind a single abstraction, using context caching to reuse each parsed proxy across every analysis batch

Ownership-weighted risk scoring via a nightly Snowflake sync of shareholder positions, so compliance flags reflect each investor’s actual stake

TECH STACK
Next.jsTypeScriptFastAPIPostgreSQLAWS SQSS3SnowflakeGeminiClaudeMistral OCR

Results

0+

API Endpoints Shipped

0

LLM Providers, One Abstraction

0+

Policy Categories Analyzed

This engagement is published under NDA. “Quorate” is a pseudonym, and the interface above was rebuilt with fictional data. The architecture, pipeline, and scope described here are real.

Client confidentiality

References available on request

Have a similar project? Let's talk.

We'd love to hear about your document-heavy AI or compliance platform challenges. Let's talk architecture.

START A PROJECT