Building self-improving tax agents with Codex
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
★ Tier-1 Source
By Members of Technical Staff: Aravind Srinivasan & Samay Shamdasani (Thrive Holdings), Arthur Fernandes Araujo & John de Wasseige (OpenAI)
Key facts
- The team measure accuracy by checking what share of returns reach 75%, 90%, or 100% correct field completion
- At launch, only a quarter of returns were at 75% correct field completion, but within six weeks, 86% hit that mark
- By Members of Technical Staff: Aravind Srinivasan & Samay Shamdasani (Thrive Holdings), Arthur Fernandes Araujo & John de Wasseige (OpenAI)
- It saves practitioners about a third of their time on tax preparation, drafts returns with up to 97% accuracy, and increases throughput by about 50%, creating more room for them to spend time
Summary
By Members of Technical Staff: Aravind Srinivasan & Samay Shamdasani (Thrive Holdings), Arthur Fernandes Araujo & John de Wasseige (OpenAI) How Thrive Holdings and OpenAI co-developed Tax AI for Crete accountants by fusing practitioner expertise with a Codex-driven loop. Real-world systems behave differently in production than they do in a lab, breaking in ways that are hard to anticipate before deployment. In this post, they'll unpack how they used Codex to build this type of agent. Crete practitioners prepare tens of thousands of tax returns each season which requires working through millions of underlying documents.