From Daily Ship Cycles to QA Bottlenecks: Solving AI Automated Testing Accuracy
No-Common1466 · reddit · 2026-08-19
A dev team using Claude Code and Codex with PRDs, mockups, and ADRs has achieved rapid shipping, sometimes finishing features within a day. However, QA has become the major bottleneck as development speed increased.
Current Struggles:
- Tried using Codex to verify and QA based on JIRA acceptance criteria with 3-5 retries.
- First-pass pass rates are very low, bug reopen rates are high, and new issues are found during testing.
- Even with AI review before merging PRs, code quality remains suboptimal.
The author seeks advice: How to improve AI QA automation? What should human QA focus on? Are there open-source tools specifically for AI code quality and QA?
More from coding & agent
- openwiki 0.3.3 launches with built-in MCP connector and tool discovery — hwchase17 · 2026-08-19
- Puck now supports real-time voice for agent coordination — HankYeomans · 2026-08-19
- Code.Storage Opens Signups: Unlimited Git Infrastructure for AI Agents — dsp_ · 2026-08-19
- Using Directed Graphs for Predictable Agent Workflows — rseroter · 2026-08-19
- Dev vibe-codes a mobile shell over Maestro, prompts local agents from iPhone via Tailscale — cocktailpeanut · 2026-08-19
- LlamaParse Outperforms General VLMs in Document Visual Grounding — llama_index · 2026-08-19