Traditional BI asks what the dashboard shows. An investigation asks what might be happening, and what evidence supports that reading. Those are different jobs. One is a report. The other is a programme of questions against data that may be incomplete, delayed or politically inconvenient.
Stack is built for that second job. You point it at organisational data and ask an open-ended question. It forms hypotheses, tests the ones that can be tested, and turns findings into a structured analysis — not a guaranteed account of the truth.
Query sits earlier in the same grammar: measure what users actually do, then let an agent investigate why the metric moved. ITAD Assist is a production example of the same instinct — compare routes against evidence rather than intuition — even though it is a bespoke system, not Stack.