Projects

What this flow shows: how a predictive model becomes something a project manager can act on. Risk enters through the portfolio view the PM already uses, narrows to a ranked list, then opens into a single project where the AI states its case: a plain-language summary, a predicted date, and the historical base rate behind it. Each risk expands into why it was flagged — the contributing factors and how often this pattern preceded a delay — and then into what to do about it, via Dela's mitigation suggestions. The principle throughout: signals, not scores. No raw probability is ever shown without the evidence and the next step attached to it.