The Current Challenges of Test Automation and Software Quality in 2026

Software is being developed faster than ever. But can your approach to quality keep up?

Modern engineering teams are under constant pressure to deliver faster. CI/CD, cloud platforms, microservices and AI-assisted development have dramatically increased the speed at which software can be built and released.

But faster development creates a difficult question:

How do you know the software you're releasing actually works?

The answer isn't simply to write more automated tests.

Organisations today are dealing with increasingly complex applications, shorter release cycles, unreliable test suites, difficult test data, AI-generated code and growing expectations around quality.

This is why effective test automation is becoming less about the tools and more about strategy, engineering and risk management.

When quality becomes a problem, it's often already a business problem

A pattern I'm increasingly seeing in growing software companies is this:

They start with a small engineering team, no dedicated QA function and a strong focus on getting the product built and into customers' hands.

And initially, that can work.

Developers test their own work. Product managers carry out some acceptance testing. Issues are fixed as they are discovered. Everyone knows the product and communication is easy.

But then the company grows.

There are more developers. More features. More integrations. More customers. More releases.

The complexity increases — but the approach to quality hasn't changed.

I've recently seen this happen with a company that started without any dedicated QA engineers around two years ago.

At the beginning, the lack of a formal QA function wasn't necessarily a problem. The priority was building the product and establishing the business.

Two years later, the situation is very different.

The product has grown significantly, development velocity has increased and more functionality is being delivered.

But increasingly, more defects are making their way into production.

Issues that previously would have been caught before release are being discovered by customers or after deployment. Regression risk is increasing, and the existing development process is no longer providing the level of confidence the business needs.

The company has reached an important realisation:

It doesn't simply need someone to write automated tests. It needs quality expertise.

That distinction is important.

Hiring a QA engineer to add tests to an existing process may help, but it doesn't necessarily address the underlying problem.

The organisation needs someone who can look across the entire delivery process and answer questions such as:

• Where are defects currently being introduced?

• What should be tested and at what level?

• Which user journeys represent the greatest business risk?

• What should be automated?

• Where is manual testing still valuable?

• How should test data be managed?

• How reliable is the existing test environment?

• How should quality be measured?

• How can testing keep pace with development?

• What needs to change in the engineering process itself?

  
  

This is where Quality Engineering and experienced QA leadership become increasingly important.

The goal isn't to put a QA gate at the end of the development process.

It's to build quality into the process from the beginning.

1. Software is being developed faster than it can be tested

The biggest challenge facing many QA teams isn't automation itself.

It's velocity.

Development teams can now produce and release changes incredibly quickly. AI coding assistants and modern development platforms are accelerating this further.

The problem is that testing doesn't always scale at the same rate.

The 2025–26 World Quality Report found that only around one-third of test cases are automated on average, despite significant investment in automation and AI.

When development velocity increases but quality processes don't evolve with it, testing becomes a bottleneck.

The objective shouldn't be to automate everything.

It should be to create a quality strategy capable of keeping pace with software delivery.

2. Automation without strategy creates technical debt

One of the most common mistakes organisations make is assuming that introducing an automation framework automatically creates an effective automation strategy.

It doesn't.

A framework is a tool.

A strategy determines:

• What should be tested

• What should be automated

• At which level it should be tested

• What should remain manual

• How tests should run

• How test data is managed

• How failures are investigated

• How quality is measured

Without this thinking, teams can end up with hundreds or thousands of automated tests that provide surprisingly little confidence.

More tests don't necessarily mean better quality.

The right tests, covering the right risks, matter far more than the size of the test suite.

3. Flaky tests destroy confidence

Few things damage an automation programme faster than flaky tests.

A test that passes sometimes and fails sometimes without a genuine product defect creates noise.

Eventually, developers start ignoring failures.

That's dangerous.

The purpose of automation is to provide fast, reliable feedback about software quality. If engineers can't distinguish a genuine regression from an unstable test, the feedback loop breaks down.

Flakiness can be caused by:

• Poor test design

• Shared or unreliable test data

• Timing and synchronisation issues

• Unstable environments

• External dependencies

• Network problems

• Race conditions

• Poor test isolation

Simply rerunning failed tests isn't the answer.

Retries can hide the underlying problem.

A mature automation strategy treats test reliability as an engineering problem and works to eliminate the causes of instability.

4. Test data remains a major automation challenge

You can have an excellent automation framework and still struggle to achieve reliable testing if your test data is poorly managed.

Automated tests frequently depend on specific customer records, account states, permissions, transactions, dates, configurations and external system responses.

Creating and maintaining that data can become one of the slowest parts of automated testing.

Effective test data management should therefore be treated as part of the automation architecture — not an afterthought.

In many environments, API-driven test data setup can dramatically reduce execution time and make tests more independent.

5. AI is changing test automation — but it isn't replacing Quality Engineers

AI is arguably the biggest change happening in software testing today.

AI-powered tools can assist with:

• Generating test cases

• Generating automation code

• Creating test data

• Analysing failures

• Identifying potential coverage gaps

• Test prioritisation

• Maintaining automation

• Generating documentation

But there is an important distinction:

AI can accelerate testing. It doesn't remove the need for engineering judgement.

AI-generated tests can still test the wrong thing.

AI-generated automation can still be poorly structured.

And AI-generated results still need to be evaluated.

The future is therefore unlikely to be "AI versus testers."

It is more likely to be Quality Engineers working with AI to achieve greater coverage, speed and insight.

6. AI-generated software creates a new quality problem

AI isn't only changing how we test software.

It's changing what we have to test.

Organisations are increasingly introducing AI-powered features, LLM integrations and agentic workflows into their products.

Traditional functional testing isn't always sufficient for these systems.

An AI system can technically function while still producing an unacceptable result.

Quality now involves questions such as:

• Is the output accurate?

• Is it consistent?

• Does it follow the intended instructions?

• What happens when the input is ambiguous?

• Can it hallucinate?

• Does it expose sensitive information?

• Does behaviour change between model versions?

• How do we evaluate probabilistic outputs?

The question is moving from:

"Does the software work?"

to:

"Does the software behave reliably under the conditions that matter?"

That requires new approaches to test design, evaluation and monitoring.

7. UI automation shouldn't carry the entire testing strategy

End-to-end UI tests are valuable because they validate real user journeys.

But they're also typically slower and more expensive to maintain than lower-level tests.

A modern automation strategy should use the right level of testing for the right purpose.

Unit tests

Fast feedback on individual components and business logic.

API and service tests

Efficient validation of business logic, services and integrations.

UI tests

Validation of critical end-to-end customer journeys.

Integration tests

Confidence that systems and services work correctly together.

Exploratory testing

Human investigation of behaviour that automated tests may not anticipate.

The goal isn't to eliminate UI automation.

It's to avoid making UI automation responsible for testing everything.

A well-designed automation strategy should provide maximum confidence with minimum unnecessary maintenance.

The shift from QA to Quality Engineering

The example of a company moving from zero dedicated QA engineers to recognising the need for quality expertise within two years is representative of a wider challenge.

Quality can no longer be something that happens immediately before release.

It needs to be considered throughout the software development lifecycle.

That means Quality Engineers working closely with:

• Product

• Engineering

• DevOps

• Architecture

• Security

• Data

Quality becomes a shared engineering responsibility rather than something handed to a QA team at the end of development.

This is the essence of Quality Engineering.

The role of an experienced QA professional is therefore changing.

It's not simply about executing test cases or building an automation framework.

It's about understanding the entire delivery system and identifying where quality can be improved.

What should organisations do?

There isn't a single automation framework or AI tool that will solve these problems.

The fundamentals still matter.

Start with business risk

Identify the workflows and capabilities that matter most to customers and the business.

Build an automation strategy

Decide what should be automated, where it should be tested and why.

Make automation reliable

Treat flaky tests, poor test data and unstable environments as engineering problems.

Use AI intelligently

Use AI to accelerate experienced engineers rather than simply generating more tests.

Improve the feedback loop

Make failures fast, meaningful and actionable.

Measure outcomes

Focus on software quality and business risk rather than simply counting automated tests.

Make quality everyone's responsibility

QA should provide the expertise, strategy and engineering capability — but quality belongs to the entire delivery organisation.

The future of test automation isn't more tests

The next phase of test automation isn't about creating the biggest possible test suite.

It's about creating better feedback with less waste.

AI will change how tests are created, maintained and executed. Automation will continue moving deeper into CI/CD. Test data will become increasingly automated and synthetic. Quality Engineering will become more closely integrated with software development.

But the fundamentals won't disappear.

You still need to understand the product.

You still need to understand risk.

You still need to understand what customers expect.

And you still need experienced engineers who can determine whether testing is actually providing meaningful confidence.

The organisations that succeed won't necessarily be the ones with the most automation.

They'll be the ones that build a quality strategy capable of keeping pace with the way software is actually being developed and delivered.

How Anchor Quality can help

At Anchor Quality, we help organisations build practical, scalable approaches to QA, test automation and Quality Engineering.

Whether you're starting an automation programme, struggling with an unreliable test suite, seeing more defects reach production, scaling an engineering team or trying to understand how AI can improve your testing strategy, the focus should always be the same:

Better quality. Faster feedback. Lower delivery risk.

That's where effective test automation delivers real business value.

Anchor Quality QA Strategy, Test Automation & Delivery Reliability