Fast, Smart, and Fragile: Rethinking Testing Leadership in the Age of AI
summary
AI-generated tests arrive 3x faster test creation and cut scripting by 70%, but a Microsoft study shows Copilot yields ~55% faster coding and speed hides gaps: convoluted suites, auto-healing that masks regressions, and blurred accountability that can scale failures like $440 million lost in 45 minutes. Leaders must keep humans doing risk-based prioritisation, require named sign-offs for release decisions, log AI decisions, set confidence thresholds that trigger review, and rotate skills to preserve critical thinking.
AI-written summary · about
chapters (44)
- 0:00Start
- 0:05Session Opens - Fast, Smart and Fragile
- 1:06Speaker Intro - Laveena Ramchandani, QE Manager, easyJet
- 2:40Mixed Feelings About AI - and Why She Argues for Embracing It
- 3:43Agenda: Acceleration, Hidden Risks, What Stays Human, Ownership, Playbook
- 4:43Testing Has Never Moved This Fast
- 5:153x Faster Test Creation, 70% Less Manual Scripting
- 5:45The Leader's Job: Knowing Which Tests Actually Matter for This Release
- 6:15Do Your Risk Assessment as Humans, Not Through AI
- 6:46Convoluted Suites of Thousands of Tests Nobody Triaged
- 7:19After AI: Describe Intent, Agents Generate Steps, Self-Heal on UI Shift
- 7:49Why Prompt Engineering Becomes the Skill That Matters
- 8:49The Microsoft Study: 55% Faster With GitHub Copilot
- 9:19Beneath the Speed: A Growing Blind Spot
- 9:49Who Chose This Test? Why Was It Marked Pass When It Failed?
- 10:21Ownership Blurs Exactly When Risk Should Be Easiest to See
- 10:53The Knight Capital Case, 2012
- 11:24440 Million Lost in 45 Minutes Before Anyone Stepped In
- 11:57AI Speeds Up Testing and Scales Risk Silently
- 12:28What Should Stay Human: Risk-Based Prioritisation
- 12:58Exploratory Testing, Accessibility and Ethical Goals
- 13:29Final Release Sign-Off Must Never Be AI's
- 14:02Auto-Healing Agents and the Question They Raise
- 14:32A Green Check Mark Is Not a Reviewed Change
- 15:02Skills Atrophy Quietly - When AI Always Answers, People Stop Asking
- 15:33Share Your Prompts - Including the Ones That Failed
- 16:03New Engineers Never Learn the Why
- 16:33Two Testers, Same AI Suggestion, Very Different Outcomes
- 17:35The Difference Is Critical Thinking - Protect It Deliberately
- 18:06Ownership in an AI-Driven Pipeline
- 18:37If AI Can Act, a Named Human Must Be Accountable
- 19:08The Root Cause Analysis Agent
- 19:40Human Review Before It Shapes a Release Decision
- 20:10AI Does Not Triage - Leadership Owns the Call
- 20:41The Framework: Judgement, Ownership, Trust, Resilience
- 21:14The Practical Playbook: Map the Decisions, Instrument for Visibility
- 21:44Set Confidence Thresholds That Trigger Human Review
- 22:15Recap: Not Everything Should Be Automated
- 22:47Strong Leadership Matters More Than Ever
- 24:25Q&A: The Biggest Leadership Mistake When Adopting AI
- 25:56Q&A: One Pre-AI Leadership Habit to Drop
- 27:29Q&A: The Biggest Opportunity - CX, Productivity, Operations or Decisions?
- 29:31Q&A: Stopping QA From Becoming the Bottleneck or the Casualty
- 31:03Give Leadership the Real Percentage You'd Ship Unverified
related talks
44mHow AI is Exposing the Existential Crisis of Software Testing
41mHow AI is transforming software testing
06mTriple Threat: AI Technologies Reduce Testing Costs
47mPanel Discussion: QE Strategic Shift: What's Changing with AI, Automation, and Speed?
29mCritical Thinking for Testers in the age of AI
42mThe Fear of Becoming Irrelevant
32mQuality engineering in the AI age: From testing AI to leveraging GenAI
01h 04mImpact of AI on #SoftwareTesting: Are Testers Ready? 🔮