Program Thursday March 4, 2027

08.30 Registration
 

Ard Kramer
09.15 Announcement & Opening
Ard Kramer

Senior Consultant, Speaker and Trainer


Pascal van der Heiden
09.30 Keynote - From Spec to Proof: The AI-Assisted Tester’s Cockpit

When AI can generate code and tests in minutes, a green pipeline is no longer enough. Are we testing the right behaviour—or just proving that the implementation agrees with itself?
This demo-driven session puts the tester at the centre of spec-driven development. Starting with an imperfect feature specification, we challenge assumptions, sharpen acceptance criteria and turn intended behaviour into automated checks.
Our workbench is a custom interactive canvas in the GitHub Copilot app: a tester’s cockpit connecting requirements, test scenarios and execution evidence. Instead of scrolling through prompts and logs, we inspect gaps, review agent-proposed tests and decide where further investigation is needed. Then we change a requirement and follow the impact through the test suite. Along the way, we combine reusable testing skills with human judgement to reduce repetitive work without automating away critical thinking. We also expose the blind spots of reassuring metrics: high code coverage does not mean the specification is satisfied.
You’ll leave with a practical pattern for connecting specifications, test automation and evidence—and a different view of tester productivity: less time assembling information, more time asking the questions that prevent defects.

Pascal van der Heiden

Senior Solution Engineer for Cloud en AI Software, Microsoft Netherlands


Irja Straus
10.10 Keynote - Roads? Where We’re Going, We Don’t Need Roads

I left testing once, for product management. There I learned what the business actually asks from testing: the confidence to ship, not our tests or a shiny green pipeline. That confidence is ours to give. Every time a new tool shows up, we in testing ask what will happen to us. The short answer is: nothing, we turn out fine! The long one is: only if we stop delegating what is ours.
Let’s take the DeLorean and look through the times I’ve tested in, from SaaS products to robots, and find out which parts of the test automation craft survived every bump on the road, and what is ours to keep. You will see that this craft isn’t something anyone takes from us. It’s something we choose to hand over, or hold on to. And wherever the road ends, you will know how to keep going.

Irja Straus

Team Lead Testing Electronic Components, Jungheinrich


10.50 Break and visit exhibition
 

11.20 Breakout sessions round 1

Richard Dekker
A1 - Same Model, Different Agent Why AI still needs a tester's mindset

AI has dramatically improved engineering practices. User stories are better, test automation is generated by default, and developers increasingly deliver features together with automated checks. Yet, despite having access to the same models, developers and testers often arrive at very different conclusions. In this talk Richard explores why that happens. Using real examples from an AI-first development environment, he shows how AI agents inherit the goals of the people directing them. Developer agents optimize for proving that software works. Tester agents optimize for discovering where it doesn’t. The result is a surprising conclusion: AI is not removing the need for testers. It is making the tester’s mindset more valuable than ever.

Richard Dekker

Software Tester, Squerist

Session for experience attendees

Presented by Squerist


Christian Baumann
B1 - Six Agents Walk Into a Test Plan: What They Handed Back

How is the AI hype treating your test automation? Barely a week goes by without a tool that claims to do your job. Read the story, write the tests, build the suite, find the bugs. All on its own. Has anyone asked you yet why the team still needs you?
I built an agentic testing workflow to find out. Six expert agents, from requirements review to automation & exploratory testing. Each goes on trial. Where does it help? Where does it break? What lands back on my desk? No tool demo, no “we found 17 bugs with AI” story. You get what I built, what I threw away & the work no agent took.

Christian Baumann

Principal Test Architect, MaibornWolff GmbH

Session for experience attendees

Joep Schuurkes Oliver Verver
C1 - We Have Some Cool Tests and We Can Show You the Code!

Our team is building the software that determines the results of elections in the Netherlands. Because transparency is crucial to safe elections, all our code is public. So not only can we talk about our tests, we can actually show them to you!

Joep Schuurkes

Staff Test Engineer, Kiesraad

Oliver Verver

Kiesraad

Session for experience attendees

Maria Shalnova-Weinzierl
D1 - Smarter UI Testing with AI: What Works, What Doesn’t

UI test automation often involves flaky tests, fragile locators, and high maintenance effort. AI promises improvements, from self-healing tests to autonomous test creation. But do all AI capabilities deliver tangible value? This talk explores practical AI-assisted UI testing, including element identification, visual validation, failure analysis, and test maintenance. Learn where deterministic automation remains essential and how to introduce AI responsibly while maintaining trust in test results.

Maria Shalnova-Weinzierl

Test Automation Engineer, ZEISS Group

Session for beginners

12.00 Change rooms
 

12.10 Breakout sessions round 2

Craig Risi
A2 - Test Data as a Service (TDaaS): Solving One of Testing’s Biggest Bottlenecks

One of the biggest hidden bottlenecks in modern software testing isn’t automation, infrastructure, or tooling – it’s data. Without reliable, compliant, and readily available test data, even the best test frameworks stall. This talk introduces Test Data as a Service (TDaaS): a scalable, self-service approach that empowers teams to provision consistent and secure test data on demand.

Craig Risi

VP Software Engineering, Entersekt

Session for beginners

João Proença
B2 - The Agent is as Good as its Harness: Test Automation in the Agentic Era

With GenAI and coding agents, a new craft has risen: harness engineering, building the environment, rules and tools that let an AI agent operate safely and accurately. Much of that harness is automated feedback, so test automation and harness engineering are intertwined, and what test automation means is shifting with it. Drawing on first-hand experience building a harness, from nudging coding agents org-wide towards TDD to the mistakes AI makes writing tests, I show where test automation fits and why it matters more, not less.

João Proença

Staff Quality Engineer, Pleo

Session for experience attendees

Ghislain Gabriëlse
C2 - Test the Factory, Not Just the Car: Quality Control for AI-Driven Software Development

Is test automation still relevant now that AI can write most of our code? Car factories faced it first. GM filled a plant with robots in the eighties, and Tesla made the same mistake in 2018. Their lesson: automation doesn’t fix your quality. It copies it, only faster. In this talk I take you inside the AI factory I build for 25+ teams at BMW Financial Services, including a platform bug that kept our skills away from the agents for months while every test stayed green. Then I show, how we test the factory now. You leave with a blueprint: the open-source framework I demo, ready to clone.

Ghislain Gabriëlse

Test Automation Consultant, deTesters

Session for experience attendees

Dr. Jan Reimann
D2 - Observability for Test Automation: Turning Test Reports into Queryable Data Before We Let AI Answer Our Questions

Test reports only answer the questions they were designed for, while the relevant questions often come up later. Before we let AI answer them, or even decide if we should, our test results need to be queryable at all. In this talk I explain how we apply observability concepts to test results: a custom Playwright reporter writes every result into OpenTelemetry and Grafana queries it. I will show real queries, e.g. how much CI compute time retries of flaky tests cost us, or which tests fail suspiciously often together, a cross-correlation the reporting tools I know don’t do.

Dr. Jan Reimann

Full Stack Quality Engineer, QualityMinds GmbH

Session for experience attendees

12.50 Lunchbreak and visit exhibition
 

Ard Kramer
13.50 Afternoon kick-off
Ard Kramer

Senior Consultant, Speaker and Trainer


Joe Colantonio Jim Trentadue
14.00 Keynote - Relevant to Whom? The Meeting Between AI Hype and Delivery Reality

You’ve sat in this meeting. The AI Center of Excellence shows up with slides full of capabilities, live demos, and big promises. The delivery lead shows up with deadlines, locked test data, a change advisory board, and a flat budget. Both walk out convinced the other one doesn’t get it.

In this keynote, Joe Colantonio and Jim Trentadue recreate that meeting live on stage. Joe plays the AI enthusiast who, by his own admission, talks to vendors all day and hasn’t had a real job in seven years. Jim plays the delivery lead who has to make it work on Monday. Both bring real data from 505 testers in the TestGuild State of Automation survey, and the data backs up each of them.

Along the way, you’ll watch AI agents build working tests, and then watch one fail with total confidence. The result is an honest answer to this year’s question, “Is test automation becoming less relevant?” It isn’t. Its value is shifting to the people who can turn what AI promises into what a team can actually deliver.

Joe Colantonio

Founder, TestGuild

Jim Trentadue

QE & Automation Leader Driving Transformation, TestGuild

Presented by Squerist

14.40 Change rooms
 

14.50 Breakout sessions round 3

Andrei Contan
A3 - Beyond the Comfort of Fast: Building Teams That Steer, Not Just Ship

We’re making software faster than ever. AI writes code, generates tests, reviews PRs. Engineering dashboards glow green, but here’s the uncomfortable question: are we shipping better products, or just more of them? This talk uses real experiment data to show that the metrics we optimize for code quality, test coverage, deploy frequency don’t predict whether our work actually succeeds. The bottleneck was never speed. It’s the missing feedback loop between “we shipped it” and “it worked for our customers.”

Andrei Contan

Principal Test Engineer, Booking.com

Session for experience attendees

Ruben van Kruistum
B3 - Still relevant: what an AI-assisted migration actually depended on

EDSN’s Robot Framework suite (320+ test cases, 10,000+ lines) had to move from SeleniumLibrary to Browser Library (Playwright). It was written off as too big to estimate: shared low-level code ruled out migrating in steps, and a months-long rewrite would chase a moving codebase. We did it in a few weeks, inside EDSN’s own approved AI environment. The LLM was the smallest part; deterministic parsing, linting, execution and human review did the rest. This talk shows why AI makes good test automation engineering more relevant, not less.

Ruben van Kruistum

Tech Lead AI Solutions, QA Company

Session for experience attendees

René van Veldhuijzen
C3 - Coverage is Dead. Long live Confidence!

This session challenges the idea that more tests automatically lead to better software quality. Drawing on real-world incidents, it introduces the Confidence Portfolio Framework: a practical model for building trust through automation, observability, monitoring, resilience, contract testing, and customer feedback. Learn how to identify confidence gaps, avoid the false security of ever-growing test suites, and redefine quality engineering in the age of AI by shifting focus from test coverage to release confidence.

René van Veldhuijzen

Strategic Consultant & Test Positivo, Squerist

Session for beginners

Petros Plakogiannis
D3 - Are Testers Still Relevant When AI Writes the Answers?

What happens to test automation when the system under test no longer gives predictable answers? In this session, we’ll explore how an AI-powered airline chatbot can be tested using Promptfoo. Through practical examples, we’ll see how familiar testing concepts evolve to cover hallucinations, inconsistent responses, safety, and quality and why testers may be more relevant than ever in an AI-powered world.

Petros Plakogiannis

Principal Quality Assurance Engineer, Quento

Session for beginners

15.30 Break and visit exhibition
 

16.00 Breakout sessions round 4

Emiel Smetsers
A4 - From Performance Testing to Performance Assurance: What 25 Years Have Taught Me, and Where AI Takes the Profession Next

2026 marks 25 years since I started in performance testing, long enough to reflect honestly on what’s worked and what hasn’t. There have been many shifts along the way; the current one is AI fast taking over performance-test scripting, so testers will need less tool- and scripting-specific knowledge going forward. What grows instead is the need to analyse results, give sound advice and, above all, make explicit what an application is actually meant to deliver. That shift moves the discipline from performance testing toward performance assurance.

Emiel Smetsers

Director Team PerformanceArchitecten, Cerios

Session for beginners

Alisa Hrustic
B4 - Long Live Test Automation: a Practical Framework for Testing LLM-based Applications

AI is becoming part of the product, and traditional test automation, built on the idea that the same input always gives the same output, can’t keep up. This talk argues that test automation is more relevant than ever, but its approach must change. A live demo presents a Python, pytest and Playwright framework that tests an LLM-based loan assistant for a bank. It shows classic tests breaking, new tests catching a hallucinating AI, and why the data the AI relies on defines what “correct” means. The result is a practical pattern for testing LLM-based applications.
AI is no longer just a tool we use to build software. It is becoming part of the product itself. Release cycles keep getting shorter, and LLM-based features now reach production in banking, insurance and customer service. Traditional test automation was built on one assumption: the same input gives the same output. LLMs break that assumption, and many existing test suites quietly become flaky or blind.

Alisa Hrustic

Test Manager, System Verification

Session for all attendees

Linda van de Vooren
C4 - Managing Up: A Test Automation Story

A test automation engineer at a bank is told to automate only what their own team builds, leaving integration with neighbouring systems to a rushed manual test before go-live. As a testlead, I saw the same gap across teams: no system-level automation, and extra engineers changed nothing. I talked to their manager, created space for the engineers to work, and an integration framework emerged. This talk shows how I convinced the organisation, and how you can convince your manager about solutions you believe in, big or small.

Linda van de Vooren

Consultant, Bartosz ICT

Session for experience attendees

Risko Ruus
D4 - Can AI See the Bug? Benchmarking Multimodal LLMs for Visual Testing

Pixel-diff tools show that a UI changed, not whether the change matters. This talk shows how multimodal LLMs can spot real visual bugs that pixel comparison misses, such as broken layouts, missing elements, and contrast issues. It introduces two open benchmarks I built: a discovery benchmark measuring how well models find unknown defects in app screenshots, and an assertion benchmark measuring how well they judge whether a claim like “the login button is in the top right corner” is true. Attendees leave with practical ideas for more accurate, scalable UI validation.

Risko Ruus

Principal QA Engineer, Rush Street Interactive

Session for experience attendees

Dorien Saliën
E4 - Testing as a Story: The Legend of Mark

What happens when “Mark knows”, but Mark has left the company? 😱 Your test suite could be the best documentation you have, if people can actually understand it. In “Testing as a Story: The Legend of Mark”, learn how to turn cryptic Cypress tests into readable, living documentation that explains how your application behaves.

Dorien Saliën

Full Stack Tester, b.ignited

Session for beginners

16.40 Change rooms
 

Janna Loeffler
16.50 Keynote - The Bots Took My Job (I'm Fine)

Twenty years ago I started my career writing test scripts. Today, AI writes them faster, runs them more often, and doesn’t complain about flaky environments. So what’s left for me? Turns out, quite a lot. This talk is about the pivot from “person who tests things” to “person who makes sure it’s safe to ship things,” a shift I call Safe Delivery Enablement. It’s less about mourning the test script and more about claiming the more interesting job that’s opening up in its place: the one where humans decide what “good enough to release” actually means, and AI just does what it’s told.
Come for the mild existential crisis, stay for the practical playbook on where quality engineers add value now that the robots do the typing.

Janna Loeffler

Director Quality Engineering, The Walt Disney Company


Ard Kramer
17.30 Wrap-up and closing
Ard Kramer

Senior Consultant, Speaker and Trainer


17.45 Drinks at the exhibition