Quality Revolution: How AI & ML Reshape Assurance
Principal Architect, QADBDeep Barot
Founder, ContextQA (host)
Where do AI and ML genuinely help software testing today? An experienced QA innovator separates hype from reality — converting manual cases to scripts, self-healing locators, and generating test data with prompting — while being candid about accuracy limits and the non-negotiable role of manual review.
Walk away knowing how to apply it
What the conversation covers
Why AI surged now: compute, training data, and capable models
AI across the QA lifecycle: requirements, automation, regression, impact analysis
Converting manual cases to scripts and auto-building page objects
Self-healing: finding an alternate locator instead of failing the test
Few-shot prompting for domain-specific and performance test data
The accuracy reality: 60–70% is good, improved with training and feedback loops
Real org blockers: cost, PII/security (mitigated by data masking), infrastructure
The ideas worth remembering
AI is assistive — treat outputs as accelerators, then review
~60–70% accuracy is realistic; raise it with training data and feedback
Manual review isn't going anywhere
Just as automation didn't take QA jobs, AI won't either — upskill
It will not take away your job — it is here to help you. You just have to know how to use it.— Anjali Kulkarni
Who you'll hear from
Anjali Kulkarni
Principal Architect, QA
Deep Barot
Founder, ContextQA (host)
See ContextQA in action
Go from watching to doing — spin up an AI agent and watch it test, self-heal, and report for you.