QA at the Point - October 2026 edition

We’re getting going again! Join us this month to hear Pax talk about Testing AI: When There’s No Single “Expected Result”. This will be recorded and hopefully streamed on Zoom too (link to come). We’ll have networking and snacks! Come for the learning, stay for your peers.

Talk abstract:

Traditional software testing often gives us a clear answer: this input should produce this output. AI changes that. The same input can produce different responses, and more than one of them may be correct—so what does “pass” actually mean?

In this session, we’ll look at practical ways to test AI-powered systems when exact expected results don’t exist. We’ll cover defining what “good” looks like, testing for consistency and hallucinations, building evaluation criteria and regression suites, and deciding what can be automated versus what still needs human judgment. We’ll also get into AI agents, including what changes when AI can take actions instead of just generate responses—and what happens when we use agents to test other agents.

No AI or machine-learning expertise required. This is a QA talk about applying the testing skills we already have to software that behaves differently than the systems we’re used to testing.