Historical Plan Reasoning
The AI assistant reasons across the full governed snapshot archive — when the plan began to drift, which work packages were serially rescheduled, how the forecast has evolved — answered from authoritative historical evidence.
Project questions that require historical reasoning are among the most valuable and the hardest to answer well. When did the schedule start slipping? Which work packages have been serially rescheduled? Is the current forecast more or less optimistic than it was three months ago? These questions require access to how the plan actually looked at specific points in the past — not just what it looks like today.
The Archive as Evidence
ProjectXL's AI assistant has access to the full governed snapshot archive — every named baseline, forecast checkpoint, status-to-date record, and what-if scenario that has been saved over the course of the project. When a user asks a question that requires historical reasoning, the assistant does not make inferences from current data fields or fall back on general project management heuristics. It reads the actual historical record and bases its answer on what the governed plan actually showed at the relevant points in time.
Comparison Anchor Resolution
The assistant resolves comparison anchors using the same snapshot inventory the comparison surface uses: current state versus the active baseline, current state versus a specific named snapshot, or any two snapshots against each other. When a question is anchored to a specific point in time — "how does the current forecast compare to what we showed at PDR?" — the assistant resolves the referenced snapshot by name and uses it as the comparison anchor for the response. The resolution basis is always stated explicitly in the response so the user can verify that the correct snapshot was used.