# Catch AI Regressions Before They Ship with AI Evals in CI/CD

DevFeed: [Catch AI Regressions Before They Ship with AI Evals in CI/CD](<https://devfeed.tech/articles/catch-ai-regressions-before-they-ship-with-ai-evals-in-ci-cd-13376.md>)

Original publisher: [Read original article](<https://www.harness.io/blog/catch-ai-regressions-before-they-ship-with-ai-evals-in-ci-cd>)

Author: Shibam Dhar

Published: 2026-09-02T00:00:00Z

Content type: article

Language: en

Sources: [Harness Blog](<https://devfeed.tech/sources/harness-blog.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Testing](<https://devfeed.tech/topics/testing.md>), [AI Development](<https://devfeed.tech/topics/ai-development.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-evals](<https://devfeed.tech/tags/ai-evals.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [production](<https://devfeed.tech/tags/production.md>), [quality](<https://devfeed.tech/tags/quality.md>), [tests](<https://devfeed.tech/tags/tests.md>)

## AI overview

Harness AI Evals uses golden datasets, response-quality metrics, and blocking quality gates in CI/CD to catch AI agent regressions before production. In an e-commerce support-agent test, an early run passed about 65% of cases, below the 70% deployment threshold, revealing incorrect, missing, or incomplete answers.

## Source excerpt

Harness AI Evals tests AI agent quality in CI/CD, using golden datasets and quality gates to catch behavioral regressions before production. | Blog