Randy Rice's Software Testing Page

10/06/2026

You don’t need more information. You need someone experienced beside you.

Today, you can find an enormous amount of software testing information for free. YouTube, articles, books, AI—and even the ISTQB syllabus itself.

So why take a live training course?

Because information and understanding aren’t the same thing.

The ISTQB syllabus tells you what you need to know. I teach you how to understand it and apply it—both in the real world and on the exam.

In my Live Virtual ISTQB Foundation Level (CTFL) course, you have direct access to me throughout the class. Ask questions. Explore the nuances. Tell me where something doesn’t make sense. Let’s talk about how these concepts actually work when you’re testing real software on a real project.

I’ve spent more than four decades in software development, testing, quality, consulting, and training. My goal isn’t simply to help you pass an exam. I want to help you become a better tester.

Upcoming live virtual classes:

October 26–30, 2026
November 30–December 4, 2026

The course includes an ISTQB CTFL exam (from ASTQB/AT*SQA) plus three months of access to my CTFL e-Learning course for continued study and exam preparation.

There’s plenty of information out there.

Sometimes what makes the difference is having someone experienced beside you.

https://mysoftwaretesting.com/ISTQB-Foundation-Software-Testing--Live-Virtual-Course_p_46.html

ISTQB Foundation Level Course in Software Testing - Public Course

10/02/2026

The best QA communities are built by people who are still willing to learn. I keep coming back to this: The strongest testing communities aren’t the ones with the loudest opinions. They’re the ones where people are still open enough to say, “I changed my mind.”

Back in 1988, when I became a QA manager, I had a lot to learn. Someone mentioned a group that met once a month that focused on software quality - The Kansas City Quality Assurance Association (KCQAA). Joining that group changed my career. It led me to QAI (The Quality Assurance Institute), where I collaborated and wrote with Bill Perry, and it was basically the single thing that launched my career in software testing.

The thing is, it takes effort to be part of a community. But it's worth it.

In this episode of The Value of Testing podcast, I talked with Greg Paskal about what really makes a QA community work, and his answer was simple - and honest. It’s not technology. It’s not the camp. It’s not whether you’re on the Selenium side, the Playwright side, the certification side, or the “just test it” side. It’s whether people are open and teachable.

That hit hard. Because once a community turns into a bubble, learning slows down. People stop listening. They start defending identity instead of improving craft. Greg called out something else I’ve seen too: A lot of “quality” can look real without actually reducing risk. Pretty reports. AI-generated output. Noise that looks like work, but doesn't actually move the needle.

But real testing still depends on judgment, method, and the willingness to think and ask good questions.

That’s what stuck with me most. The best communities don’t just give you answers. They sharpen you. And sometimes they help you through the seasons when you need it most - out of work, stuck in a rut, or just trying to stay current.

If you don’t have a testing community, start small. One person. One lunch. One conversation. That might be enough to change your trajectory. Worth thinking about.

I know I'm going to try to start a community of practice.

Hear the full conversation:

The strongest QA communities don’t just share tools - they sharpen ...

09/29/2026

AI can generate a long list of test cases in seconds. But can it tell you which tests matter most?

That still takes judgment: understanding the risks, spotting gaps in the requirements, choosing useful test data, and designing tests that reveal problems.

Those are the skills I focus on in my Test Planning and Design e-Learning course. The course uses a realistic retail case study and practical exercises, including test planning, equivalence partitioning, boundary values, and decision tables.

If you want to strengthen the thinking behind your tests—or help your team do the same—take a look:

Learn practical software test planning and design techniques with real-world exercises and personal instructor support from Randy Rice.

09/24/2026

How do you test AI when there is no “right” answer?

For most of my career in software testing, expected results have been fundamental.

Enter X. Expect Y. Compare the actual result. Pass or fail.

But AI changes things.

Two different answers might both be acceptable. An answer can be factually correct but irrelevant. It can be helpful but unsupported. And perhaps most dangerously, it can be completely wrong while sounding perfectly reasonable.

In this week's episode of The Value of Testing, I explore what this means for testers—and for the organizations deploying AI.

I also look at the Zillow Offers case study. Zillow's difficulty forecasting future home prices eventually contributed to hundreds of millions of dollars in inventory write-downs and the business's shutdown.

That raises a fascinating quality question: A model may be accurate—but is it accurate enough for the decisions you're asking it to make?

A small error may be inconsequential in one application and catastrophic in another. Context, scale, consequences, and risk all matter.

The test oracle problem isn't new. Experienced testers have wrestled with incomplete, ambiguous, and multiple acceptable outcomes for decades. AI doesn't create that problem—but it dramatically amplifies it.

The question isn't merely, “What testing technique should we use in this context?” It becomes: How accurate must a probabilistic prediction be before we're willing to put hundreds of millions—or billions—of dollars behind it?

Now we're talking about business context, risk appetite, economic consequences, monitoring, changing environmental conditions, model performance, organizational accountability, and governance.

We also look at:

• Expected results vs. acceptable results
• The test oracle problem
• Risk-based testing for AI
• Repeated testing and variability
• Human judgment as part of the oracle
• Drift and changing real-world conditions
• AI regression testing
• The connection between testing, monitoring, and AI governance

My main takeaway: When there is no single “right” answer, testing doesn't become less important. It becomes more important.

Because someone still has to decide what is acceptable, how we'll measure it, and how much risk we're willing to take.

🎙️ Testing AI – When There Is No “Right” Answer - https://youtu.be/9BJGucnA2yI



How do you test a system when there is no single “right” answer?T...

09/18/2026

What does a cheese pizza have to do with AI governance?

Last Friday, I picked up my usual order—a large, thin-crust pepperoni and sausage pizza. At least, that’s what I thought was in the box.

When I arrived home and opened it, I found a plain cheese pizza. I had trusted the name on the order, the person handing it to me, and my own assumption that the contents were correct.

But it doesn’t matter what they say is in the box—or what we think is in the box. What matters is what is really in the box.

That experience became the starting point for my latest episode of The Value of Testing:

“What’s Really in the Box—and What That Means for AI”

This comes at a time when AI dominates the mass media. We hear serious warnings from industry leaders calling for stronger safeguards and regulation.

We also hear sensational predictions, political positioning, corporate self-interest, and claims that concerns about AI are exaggerated—or even a hoax.

How do we separate legitimate risk from hype and misinformation?

We open the box. We investigate the evidence. We examine what these systems actually do—not simply what vendors, lawmakers, journalists, or technology leaders tell us they do.

In this episode, I discuss:

• Why traditional testing is inadequate for agentic AI

• Why an agent’s operating environment and boundaries matter

• The difference between regulation, compliance, and governance

• Why organizations need their own AI strategies, policies, controls, and accountability

• Why observation and continuous monitoring are essential

• What my experience with Y2K taught me about risk, prevention, and public misunderstanding

I was there during Y2K. Heck, I even helped cause Y2K! The "defective" code was real. We found it, fixed it, and tested the corrections. The absence of widespread catastrophe did not prove that Y2K was a hoax. It showed that remediation worked.

We need that same evidence-based thinking about AI—not panic, blind optimism, or automatic trust.

A chatbot can provide a wrong answer. An AI agent can act on that answer—and affect other systems before anyone notices.

We may never test an AI agent into certainty. But we can challenge it, constrain it, observe it, monitor it, and ensure that humans can intervene.

Trust is not a test strategy—and with agentic AI, it cannot be a governance strategy either.

Watch the episode here: https://youtu.be/Ha7fgb0h6B4

I also briefly introduce Rice Consulting’s AI Governance and Quality Assessment, designed to help organizations discover what is really “inside the box” when it comes to their use of AI.

What does a disappointing cheese pizza have to do with artificial i...

09/11/2026

Who Owns AI Quality? The Question Most Organizations Can’t Answer

As organizations rapidly adopt AI, an essential question is often left unanswered: Who is accountable for its quality?

Is it the developers, testers, data scientists, vendors, product owners, risk managers—or everyone? Who defines “good enough”? Who tests for hallucinations, bias, drift, instability, and unsafe behavior? Who monitors the system after it is deployed? And...who can turn it off if needed?

In this episode of The Value of Testing, I explore why “everyone is responsible” can ultimately mean that nobody is accountable—and the governance questions every leader should be able to answer.

Before beginning this episode, I also pause to remember September 11, 2001. This episode is dedicated to the memory of Anna Allison, a friend and member of the software quality community who was aboard American Airlines Flight 11. I remain grateful that our paths crossed, and I remember her today.

Watch here: https://youtu.be/fOTcfoXkiWI

Who owns AI quality?Is it the developers, testers, data scientist...

09/10/2026

Are you ready to take your software testing leadership skills to the next level?

I’ll be teaching a live virtual ISTQB Advanced Test Management course:

October 19–22, 2026
9:00 a.m.–5:00 p.m. EDT

This course goes well beyond managing test activities. We’ll explore how effective test managers:

• Align testing with organizational and project goals
• Develop risk-based test strategies
• Estimate, plan, monitor, and control testing
• Communicate meaningful information to stakeholders
• Build and lead effective test teams
• Improve the testing process and demonstrate its value

The course is designed for experienced testers, test leads, QA managers, and others who want to strengthen their test management skills while preparing for the ISTQB Advanced Test Management certification exam.

You’ll also benefit from practical examples and lessons I’ve gained from more than 40 years of leading, assessing, and improving testing organizations.

Interested in attending—or enrolling members of your team? Just visit https://mysoftwaretesting.com/ISTQB-Advanced-Test-Manager-Course--Live-Virtual-Course_p_117.html

ISTQB Advanced Test Manager Course - Public Course

09/01/2026

Today I’m launching my newest e-Learning course: Test Planning and Design

Good testing doesn’t begin with executing test cases. It starts with understanding what matters, where the risks are, and how to design tests that produce meaningful information.

This practical, self-paced course helps testers develop those skills through a realistic retail case study involving multiple e-commerce sites spanning multiple countries.

Participants will learn how to:

• Develop a test strategy

• Develop a practical test plan

• Identify meaningful test conditions

• Apply a variety of test design techniques such as equivalence partitioning and boundary value analysis

• Use decision tables to address complex business rules

• Recognize ambiguity and gaps in requirements

• Apply risk analysis to prioritize tests

• Consider quality characteristics beyond functionality

• Make thoughtful decisions about test automation

• Know which tests make good regression tests

Throughout the course, you’ll apply the concepts - not simply read about them.

I created Test Planning and Design for testers who want to move beyond mechanically following scripts and become more thoughtful, capable, and valuable members of their teams. It promotes the idea of high-value test assets that create value beyond testing.

The course is available today at a reduced price to celebrate the launch:

https://mysoftwaretesting.com/Test-Planning-and-Design_p_149.html

If someone on your team could benefit from stronger test planning and design skills, I would appreciate you sharing this with them.

Learn practical software test planning and design techniques with real-world exercises and personal instructor support from Randy Rice.

08/28/2026

Have we forgotten what quality really means?

While watching a documentary about Boeing, "Freefall" on Netflix, I was struck not only by the technical failures, but by what can happen when an organization ignores - or even punishes - the people who raise concerns.

At the same time, I've been cleaning my attic and found a QAI conference notebook from 1991. I gave my second conference presentation that year.

As I revisited this massive software quality conference notebook from 1991, it was interesting to remember that more than 30 years ago, quality leaders were already teaching that quality depends on leadership, systems thinking, trust, and the freedom to tell the truth.

Today, we have better tools, more automation, faster delivery pipelines, and enormous amounts of data. But have we lost some of that organizational wisdom?

In this episode of The Value of Testing, I explore:

• What we understood about quality in 1991 but seem to have forgotten today

• Why quality cannot simply be tested into a product

• The difference between an organization’s stated culture and its real culture

• What happens when employees are afraid to raise concerns

• Why dashboards, automation, and passing tests cannot replace quality leadership

• How organizations can create a culture where difficult truths are heard

• What happens when there is no accountability for poor quality

An organization’s quality culture is not defined by what leaders say when everything is going well. It is revealed by what happens to the person who brings them bad news.

You can view the podcast here: https://youtu.be/51-NSt6_wwc

08/21/2026

“All the tests passed.”

That sounds reassuring, but what does it actually tell us?

A passing test suite doesn't prove a product is ready for release. It only provides evidence about the specific conditions, data, environments, and risks that were tested. If the tests are incomplete, outdated, or focused on the wrong things, a green dashboard can create dangerous false confidence.

In this episode of The Value of Testing, I explain why test status alone is not enough - and what meaningful testing information looks like.

You’ll learn:

* What a passing test actually proves—and what it doesn’t
* Why test counts and pass rates can mislead leaders
* How weak tests can produce strong but unjustified confidence
* What leaders should ask before approving a release
* How testers can communicate coverage, risk, uncertainty, and business impact
* Why a clear recommendation is more valuable than a wall of statistics
* Should testers give a "Go/No Go" recommendation?

Leaders don’t simply need more testing data. They need decision support: what we know, what we don’t know, what could happen, and what the test team recommends.

Testing cannot guarantee that nothing will go wrong. But when test results are translated into meaningful information, they can give leaders a sound and defensible basis for deciding what to do next.

Check out the podcast at https://youtu.be/DaI388swvGM

“All the tests passed.”That sounds reassuring - but what does it ...