# QA and test automation in 2026: trends and challenges

URL: https://sdclabs.com/qa-test-automation-challenges/

> In 2026, 89% of organizations pilot or deploy generative AI in quality engineering, but only 15% have scaled it, while flaky tests are becoming more common.

# Faster delivery needs stronger tests: QA and test automation in 2026

[Mykhailo Poliarush](https://sdclabs.com/author/poliarush/) 2026-11-06Updated 2026-11-175 min read

In 2026, QA and test automation are where the AI shift gets checked: AI drafts more code and more tests, and automated tests decide what is safe to ship. In the [World Quality Report 2025](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) , 89% of organizations were piloting or deploying generative AI in quality engineering, but only 15% had reached enterprise-wide implementation. Most teams are still working out how to make AI-assisted testing reliable.

## How far has AI gotten into QA?

Into most teams, but not deep into most of them. The [World Quality Report 2025](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) , a survey of more than 2,000 senior executives in 22 countries, found 37% of organizations using generative AI in quality engineering in production and 52% still in pilots. Only 15% had [enterprise-wide implementation](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) , 30% used it in limited use cases and 43% were still experimenting. The report is published by Capgemini, Sogeti and OpenText, which sell quality engineering services and tools, so treat it as vendor research.

Generative AI in quality engineering: how far organizations have got

**Enterprise-wide implementation**

15%

**Limited use cases**

30%

**Experimental phase**

43%

**Not adopting**

11%

Source: [Capgemini, Sogeti, OpenText, World Quality Report 2025 (2,000+ executives, Nov 2025; vendor)](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/)

Developers are split on handing testing to AI. In [DORA's 2025 report](https://services.google.com/fh/files/misc/2025_state_of_ai_assisted_software_development.pdf) , 62% of respondents who create test cases use AI to help. Yet in the [Stack Overflow Developer Survey 2025](https://survey.stackoverflow.co/2025/ai) , 44.1% of respondents said they do not plan to use AI for testing code. The [World Quality Report](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) adds that test case design and requirements refinement now lead AI use in QA, and that organizations report an average productivity boost of 19%, while one third see minimal gains.

## Why does AI-assisted delivery need more testing, not less?

Because AI makes it faster to produce changes, and each change can break something. The [DORA 2025 report](https://blog.google/innovation-and-ai/technology/developers-tools/dora-report-2025/) found that AI adoption now improves delivery throughput but still increases delivery instability. In the [Stack Overflow Developer Survey 2025](https://survey.stackoverflow.co/2025/ai) , 66% of developers named "AI solutions that are almost right, but not quite" as a frustration. And in [Veracode's 2025 GenAI Code Security Report](https://www.veracode.com/press-release/ai-generated-code-poses-major-security-risks-in-nearly-half-of-all-development-tasks-veracode-research-reveals/) , AI-generated code introduced security vulnerabilities in 45% of the tested cases (Veracode sells application security tools, and the test used 80 curated tasks).

More code, written faster, with errors that look plausible: that is a testing problem. Automated tests are the cheapest way to check every change at the speed AI produces them.

## What are the main challenges for QA teams in 2026?

- **Flaky tests are getting more common.** In [Bitrise Mobile Insights 2025](https://www.bitrise.io/blog/post/bitrise-mobile-insights-report-defines-new-benchmarks-for-app-velocity-and-performance) , based on more than 10 million builds from January 2022 to June 2025, the share of mobile teams that experienced any test flakiness grew from 10% to 26%. A suite that fails at random trains people to ignore red builds.
- **AI tools bring new risks.** The top challenges that World Quality Report respondents faced with generative AI in QA were data privacy risks (67%), integration complexity (64%) and hallucination and reliability concerns (60%), according to the [2025 report](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) .
- **Skills lag the tools.** In the [same report](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) , 50% said their organization lacks AI and machine learning expertise, unchanged from 2024.
- **Adoption is uneven.** The share of organizations not using generative AI in QA at all rose to 11%, from 4% in 2024, according to the [World Quality Report 2025](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) . The report reads this as the initial rush giving way to a more grounded strategy.

Mobile teams experiencing any flaky test

**2022**

10%

**2025**

26%

Source: [Bitrise Mobile Insights 2025 (10M+ builds, Jan 2022 to Jun 2025; vendor)](https://www.bitrise.io/blog/post/bitrise-mobile-insights-report-defines-new-benchmarks-for-app-velocity-and-performance)

Bitrise sells mobile CI, and its figure is a yes or no per team on its own platform, not a share of tests. Still, it matches what most engineers see: as suites grow and run more often, unreliable tests become a tax on every release.

## What does a quality failure cost?

Sometimes, a global outage. On 19 July 2024, a faulty CrowdStrike content update affected an estimated 8.5 million Windows devices, [according to Microsoft](https://blogs.microsoft.com/blog/2024/07/20/helping-our-customers-through-the-crowdstrike-outage/) . CrowdStrike's [preliminary post-incident review](https://www.crowdstrike.com/en-us/blog/falcon-content-update-preliminary-post-incident-report/) reads like a testing checklist. A bug in its Content Validator let problematic content pass validation, and the fixes it listed include local developer testing, update and rollback testing, stress testing, fuzzing and fault injection, and a staggered rollout that starts with a canary deployment.

The aggregate bill is large but hard to measure. The Consortium for Information and Software Quality estimated the cost of poor software quality in the US at [at least $2.41 trillion](https://www.it-cisq.org/the-cost-of-poor-quality-software-in-the-us-a-2022-report/) in its 2022 report, the latest edition we found. Treat it as an order of magnitude, not a precise figure.

## What does good QA and test automation look like in 2026?

The practices are not new. What changed is that they now have to keep up with AI-generated changes.

1. **Make tests the gate for every change.** Fast unit and API tests on every commit, end-to-end tests on critical user journeys, and no merge on a red build.
1. **Treat flakiness as a bug.** Track flaky tests, quarantine them with an owner and a deadline, and fix the cause: timing, shared state, test data or the environment.
1. **Let AI draft tests and people review them.** AI is good at producing test cases and data. Engineers decide what is worth testing and whether an assertion checks the right thing.
1. **Write tests as specifications for AI agents.** When an agent writes code, the tests define what done means. A clear acceptance test is a better prompt than a long description.
1. **Test the release, not only the code.** Validate configuration and content updates, roll out in stages, and keep rollback tested.
1. **Measure what matters.** Track escaped defects, flaky rate and time to feedback, not the raw number of tests.

## Who should you trust with your test automation?

People who have built the tools other teams use. SDCLabs founded the open-source testing frameworks Codeception for PHP, with 4,853 stars on [GitHub](https://github.com/Codeception/Codeception) , and CodeceptJS for JavaScript, with 4,253 stars on [GitHub](https://github.com/codeceptjs/CodeceptJS) , both as of 17 November 2026. Our senior engineers run AI-assisted workflows in which AI agents draft code and tests, engineers review and decide what ships, and automated tests verify every change.

If your suite is slow, flaky or missing where it matters, see our [QA and test automation](https://sdclabs.com/services/qa-and-test-automation/) service, the [CodeceptJS](https://sdclabs.com/test-automation/codeceptjs/) and [Codeception](https://sdclabs.com/test-automation/codeception/) pages, or our [trainings](https://sdclabs.com/trainings/) . Engagements start in 1 to 2 weeks. [Tell us about your test suite](https://sdclabs.com/contact/) .

For the pipeline side, see [DevOps and TestOps](https://sdclabs.com/services/devops-and-testops/) , and for the bigger picture, [AI in the software development lifecycle](https://sdclabs.com/ai-sdlc-trends-2026/) .

Sources checked on 2026-11-17.

## Frequently asked questions

### How widely is AI used in QA in 2026?

In the [World Quality Report 2025](https://www.capgemini.com/news/press-releases/world-quality-report-2025-ai-adoption-surges-in-quality-engineering-but-enterprise-level-scaling-remains-elusive/) , 89% of organizations were piloting or deploying generative AI in quality engineering, 37% in production, but only 15% had enterprise-wide implementation. The report is vendor research.

### Are flaky tests getting worse?

On Bitrise's mobile CI platform, the share of teams experiencing any test flakiness grew from 10% in 2022 to 26% in 2025, per [Bitrise Mobile Insights 2025](https://www.bitrise.io/blog/post/bitrise-mobile-insights-report-defines-new-benchmarks-for-app-velocity-and-performance) .

### Does AI-generated code need more testing?

Yes. [DORA 2025](https://blog.google/innovation-and-ai/technology/developers-tools/dora-report-2025/) found AI adoption still increases delivery instability, and [Veracode](https://www.veracode.com/press-release/ai-generated-code-poses-major-security-risks-in-nearly-half-of-all-development-tasks-veracode-research-reveals/) found security vulnerabilities in 45% of tested AI-generated code cases.

### Who founded Codeception and CodeceptJS?

SDCLabs founded both open-source testing frameworks. Codeception had 4,853 GitHub stars and CodeceptJS 4,253 as of 17 November 2026. See [QA and test automation](https://sdclabs.com/services/qa-and-test-automation/) .

[#test-automation](https://sdclabs.com/tag/test-automation/) [#qa](https://sdclabs.com/tag/qa/) [#flaky-tests](https://sdclabs.com/tag/flaky-tests/) [#quality-engineering](https://sdclabs.com/tag/quality-engineering/)

[Mykhailo PoliarushFounder and CEO](https://sdclabs.com/author/poliarush/)

## Related posts

[
### AI grades its own homework: why demand for QA is about to come back

Teams ship more AI-written code and understand less of it, and the same model often writes the tests. DORA links AI to delivery instability. QA closes the gap.

Quality engineering · 2026-11-17

](https://sdclabs.com/ai-grades-its-own-homework-qa-demand/) [
### Can you trust what your AI product says? Testing AI in 2026

AI testing in 2026 means measuring how often your product is wrong: hallucination rates across 26 top models range from 22% to 94% on one benchmark.

Quality engineering · 2026-11-05

](https://sdclabs.com/ai-testing-challenges/) [
### What does automation miss? Manual and exploratory testing in 2026

95.9% of top home pages fail detectable WCAG checks, and scanners find only part of the problem. Manual and exploratory testing cover what scripts miss.

Quality engineering · 2026-11-02

](https://sdclabs.com/manual-exploratory-testing/)

## Get an estimate for your product.

[Get an estimate](https://sdclabs.com/contact/) [Book a call](https://sdclabs.com/contact/#call)
