Summary
October brought improvements in the AI Advisor, per-entity cost metrics, saved custom Diagram views, a more interactive animation viewer, and a round of What-if Analysis improvements.
AI Advisor
AI Advisor can be found before you convert a diagram to a model (it can create the model for you from a blank canvas), in the Quodsi panel to provide assistance with model building, editing and updating, and in the output screens to provide assistance with interpretting results and understanding what scenarios to run to support your decision making.
Cost per entity
Results now show what each entity type costs per unit, so you can compare scenarios directly. For example, Scenario B serves walk-ins for $37 each and Scenario A for $42. You no longer need to divide total cost by throughput by hand.
Direct cost per entity is the cost of the work an entity actually used: activity costs, resource seize costs, and resource hourly costs while the entity held the resource. Idle and fixed costs are left out, so this number moves when the process itself changes.
Fully loaded cost per entity is this entity type's share of everything the run cost. It adds a share of activity fixed cost, activity idle cost, and resource idle cost. The shares for all entity types add up to the scenario's Total cost.
Read the two together. The gap between Fully loaded and Direct is what spare capacity and overhead cost per unit. Adding an idle staff member, for example, leaves Direct cost flat and raises Fully loaded cost.
Both metrics count any entity that left the model, whether it finished, was consumed by a join, was split, or was disposed. You'll find them in the entity CSV outputs, the Excel entity sheet, dashboard cards, and Tradeoff analysis axes. They stay blank if the model has no costing set up. Results from earlier runs won't show them, so re-run the scenario.
Saved custom Diagram views
If you built a custom Diagram view and then clicked a preset like Flow or Capacity, your custom view used to disappear. Now Quodsi keeps it. The Custom pill is a button that brings your view back exactly as you left it, including the measures on each shape type, the connector setting, values on lines, and any fill colors you chose. If you start from a preset and change it, that becomes your new custom view. Saved setups from earlier releases load unchanged.
There's also a new Paths preset, a connector heat map with no numbers. It shows only how often entities travel each connector, with busier connectors drawn thicker. Flow is still the default view, and Paths is one click away.
Animation viewer
Click an activity, resource, generator, or connector to inspect it. It opens the same inspector as the Diagram report, and Run stats steps aside while it's open.
A new "Go to time" box jumps to any point in the model's clock units.
Keyboard shortcuts: Space to play or pause, arrow keys to step, R to replay.
The scenario name (and replication) now shows on the stage.
Entities line up with the flow direction, and queues stack away from the step.
Step labels in drawio no longer truncate.
The "HUD" toggle is now called "Run stats".
What-if Analysis
"What-if runs" is now What-if Analysis, with a larger launch button.
Each study scenario is a full row in the Runs list. Run, cancel, view results, or delete each one individually.
A failed scenario shows its error right on its row.
The Levers tab hides settings that don't apply to your model.
Removing a column from a sentence card now rebuilds its rows correctly.
"Edit setting" works on a single element.
"Nothing yet" now reads "Baseline only".
A study sweep's default range starts one step above the current value.
New charts
Three new charts are available. Re-run your scenario to populate them.
Cost over time, with eight per-period cost series.
Generator arrivals over time, plus a per-entity creation trace.
Served vs. waiting by entity type at each activity.
Fixes and improvements
Report downloads no longer fail when model or scenario names include non-English characters.
Delete confirmations ask once, and you can't delete a scenario while its study is starting.
Reruns keep their diagram background, including a study's Baseline copy.
Editor panels no longer repeat the shape type as a subtitle.
Long runs use less memory. Results are identical to the previous release.
