# How AI Agents Saved Pricer’s QA Process > From eight QA engineers to two, Pricer still grew ESL coverage with AI agents–fewer production bugs, faster releases, and about 390 hours of testing saved every quarter. Source: https://qa.tech/case-studies/how-pricer-transform-qa-with-qa-tech · Published: 2026-03-05 --- **Company:** Enterprise B2B (Hardware + Software)**‍** **Manual QA testers**: 2 **Integration:** GitHub **Time saved**: 390h per quarter **Key Use Cases** - [Automated regression testing](https://qa.tech/use-cases/automated-regression-testing) - [Continuous smoke testing](https://qa.tech/use-cases/smoke-testing) - [Pull request & CI/CD testing (GitHub)](https://qa.tech/use-cases/pr-cicd-testing) - [AI-assisted test generation](https://qa.tech/use-cases/ai-test-generation) - [UI-change-resilient testing](https://qa.tech/use-cases/ui-change-resilient-testing) **Before**: 8 manual testers, blocked releases, and bugs in production. **After**: 2 manual testers, smooth release flow, and less bugs in production. ### Company Overview [**Pricer**](https://www.pricer.com/) builds and operates Electronic Shelf Labels (ESL) used in over 28,000 stores across 70+ countries. Their platform connects hardware and software in real-time, powering price updates and in-store coordination. The system is business-critical, and any issues cause large damage to operations. ‍ ‍ ### The problem before QA tech Pricer had a QA team of 8. Most testing done manually, with a growing suite of E2E tests. However, this did not prevent bugs from entering the production. After a company restructuring, only 2 QA engineers remained. Key issues: - Automated tests weren’t catching regressions reliably. - Test coverage was limited and brittle. - Manual QA was slowing releases down. - The platform’s complexity was larger than the QA team’s bandwidth. Pricer’s QA team needed a way to test better, not just more. ### What changed with QA tech Chris Chalkitis (CTO) pushed for a shift toward AI-driven testing and introduced **QA.tech** into the workflow — an agentic testing platform that uses AI to generate and maintain tests. Instead of scaling headcount, they used QA agents to scale test depth and frequency. > **“We project that we will significantly improve quality while saving money in the next year by investing in agentic testing that allows us to test deeper and more frequently than manual testing.”** > – Chris Chalkitis, Chief Digital Officer ‍ ![__wf_reserved_inherit](https://qa.tech/cms-assets/68de28d9c55d8552a19ec2bc_digital-price-tags.jpg) ### How it works now Each product team at Pricer owns its test plan. QA.tech agents create, update, and run test cases automatically, integrating with [GitHub to run checks on every pull request](https://qa.tech/use-cases/pr-cicd-testing). - GitHub + CI integration ensures that all critical paths are tested pre-merge. - [Smoke tests run continuously](https://qa.tech/use-cases/smoke-testing) to catch regressions early. - [Agents adapt to UI/API changes](https://qa.tech/use-cases/ui-change-resilient-testing) and propose updates to stale tests. > **“With QA.tech, we automate cases we previously didn’t have the resources to cover. Since implementation, we have increased our test coverage and significantly reduced deployment risks.”** – Thomas, Platform Development Manager ### Results - **Test coverage increased** despite the team shrinking from 8 to 2. - **Fewer bugs** in production. - **Release cycles sped up**, with fewer blockers from QA. - Engineers spend less time setting up tests and more time on features. - Teams feel more confident deploying changes. ‍ ![__wf_reserved_inherit](https://qa.tech/cms-assets/68de28ea1b3221e4f220dc83_drive-operational-efficiency.jpeg) ‍ ## Related reading - [How JOIN automated QA with AI agents](https://qa.tech/case-studies/how-join-automated-qa-with-ai-agents) - [Smartlinx replaced legacy QA processes with AI agents](https://qa.tech/case-studies/smartlinx-ai-qa-replacement) - [Dynamic PR Testing with AI Agents](https://qa.tech/product/pr-testing)