1
Step 1
Open the What-if Analysis tab
ResultsBaseline
Start from the Results page of your current scenario. Click the What-if Analysis tab to open the workspace and begin testing alternative changes based on your baseline.
AI features How-To
How to use What-if Analysis
Test controlled alternatives by changing selected assumptions, constraints, or inputs, then rerun the scenario to see how the outcome changes.
Guide details
Audience
AI feature users
Time
4 minutes
Category
AI features
What this guide covers
This guide shows you how to use What-if Analysis after results are generated. You will learn what What-if Analysis is, why to use it, what you can test, how it works, and how to read the impact of different alternatives.
Understand what What-if Analysis is for
Know why to use it
Learn what you can test
Follow how it works
Review how to read the impact
Why use What-if?
Why use What-if?
Use What-if Analysis to test possible decisions before applying them in real operations. It helps you understand how changes in timing, demand, priorities, or capacity may affect the final outcome.
What-if Analysis helps you:
Test decisions before applying
Evaluate alternative choices safely before implementing them in real operations.
Understand the impact of change
See how timing, demand, priorities, or capacity changes affect the final result.
Compare alternatives
Compare different options side by side using consistent KPIs and outputs.
Reduce uncertainty and risk
Make more confident decisions with evidence from scenario results.
Improve operational performance
Find better trade-offs to improve service, efficiency, cost, and resource use.
Support data-driven decisions
Use results, not assumptions, to choose the best path forward.
What can you test?
What can you test?
What-if Analysis lets you test changes to assumptions, constraints, and inputs across optimization scenarios.
You can explore how changes affect schedules, routes, workloads, resource usage, priorities, and overall performance before deciding which plan to use.
Available What-if options depend on the selected product and the type of optimization scenario you are running.
01
Time & Availability
Test changes related to timing, availability, operating windows, or when work can start and finish.
Examples
- Change maintenance windows
- Adjust working hours or shifts
- Modify release dates or due dates
- Change service or delivery time windows
- Update resource availability
02
Workload & Demand
Test how the plan responds when the amount of work, demand, or required activity changes.
Examples
- Add or remove jobs, orders, or stops
- Increase or decrease demand
- Change processing, service, or handling times
- Modify order quantities or workload volume
- Test higher workload pressure
03
Priorities & Rules
Adjust priorities, rules, or constraints to see how they influence the optimized outcome.
Examples
- Change job, order, or customer priorities
- Give higher weight to urgent work
- Modify due dates or service targets
- Adjust fairness, overtime, or assignment rules
- Change eligibility, skill, or routing constraints
04
Capacity & Resources
Test the impact of adding, removing, or limiting the resources available to the optimization model.
Examples
- Add or remove machines, vehicles, rooms, or teams
- Increase or reduce equipment capacity
- Change employee, driver, or crew availability
- Modify maximum route, shift, or operating limits
- Test resource shortages or extra capacity
Explore different scenarios and compare the results to choose the plan that performs best for your operation.
Baseline
Run the current scenario.
Change assumptions
Adjust inputs, constraints, or settings.
Re-run optimization
Run the model again with the changes.
Compare impact
Analyze and compare the results side by side.
Step by step
Follow the What-if path
Use these steps to start from a baseline result, describe one controlled change, review the proposed variant, run it, and compare the outcome before deciding what to keep.
2
Step 2
Review the baseline and supported changes
BaselineSupported changes
After opening the What-if Analysis workspace, review the baseline information at the top of the page to understand the current scenario. Then check the supported changes section to see what types of adjustments the system can propose or support for this product.
3
Step 3
Enter the change you want to test
PromptGenerate variant
Describe the change you want to test against the current baseline. You can type your own request in the prompt box or use one of the suggested examples to get started. Then click Generate variant.
4
Step 4
Review the proposal and run the variant
ProposalRun variant
Review the generated proposal before execution. Check the proposed changes, rationale, expected impact, and variant review. When everything looks right, click Run variant to execute the new branch result.
5
Step 5
Compare the baseline and the variant
ComparisonDecision
Review the comparison between the baseline and the variant. Check the key metrics, comparison summary, and final options to understand what changed and decide which result you want to keep.
Tips
Test one clear change at a time when possible.
Keep the baseline result unchanged so comparison stays reliable.
Use the comparison before choosing whether to rerun, keep the baseline, or save the variant as a new scenario.
Related how-to guides
Continue with nearby tasks in the same category.
How to use AI Explain
Ask focused questions about generated results so MAG OptiAI can explain KPIs, charts, schedules, routes, model outputs, or document answers using visible output.
How to use AI Configure
This guide helps you configure a scenario with confidence. You will learn what AI Config is for, how to review required inputs, how to choose configuration options, and how to validate setup before running the model.
How to use AI Compare
Compare two results side by side to understand what changed between them and decide which option performs better.
Next
Keep moving through MAG OptiAI resources
Return to the full how-to library or step back to the resource hub when you want a different kind of guide.