Clarity at the rule level
Find policy conflicts before they become real problems.
Understand how policies behave, simulate customer journeys, and uncover contradictions through traceable, deterministic analysis.
Explore the NovaCart demo. No account required.
Policy language
Original source clauses
Structured rules
LLM-assisted extraction
Deterministic analysis
Python evaluates every decision
One journey. Two policy clauses.
A replacement is cancelled.
The refund window has closed.
What should happen next? Make the interaction between rules visible, then investigate it in the workspace.
- DAY 00Item delivered
- DAY 00Replacement requested
- DAY 00Replacement cancelled
- DAY 31Refund requested
Illustrative journey using existing NovaCart rules. This static preview does not run analysis.
Grounded in the source
Refunds denied after 30 days
Denies refund“Refund requests submitted more than 30 days after delivery are strictly denied.”
NovaCart Returns Policy v1.0
Refund eligibility upon cancelled replacement
Allows refund“If a replacement order is cancelled, the customer may request a refund.”
NovaCart Returns Policy v1.0
Evidence over assumptions
Know why a decision happens.
Policy language can be ambiguous. Its behavior shouldn’t be a mystery. Explore the working tools in the current demo.
Find the disagreement
Investigate contradictions, gaps, and other findings produced by deterministic Python code.
Available in the workspace
Test the customer journey
Run the sample suite, adjust event timing, and see how the policy responds.
Available in the workspace
Follow the evidence
Inspect fired rules, source quotations, and the state after each simulated event.
Available in the workspace
Policy comparison, with shared test input
Compare demo rule sets or an explicit active-policy snapshot in What-if. Natural-language extraction is opt-in through the Analyze workspace.
How it works
From a rule to a reason.
The LLM helps structure and explain.
The deterministic engine produces findings.
- 01
Read the rules
Review NovaCart’s existing refund and replacement rules alongside their source text.
Browse policies - 02
Simulate a journey
Run sample journeys. Change the timing of an event and inspect the resulting decision.
Explore simulations - 03
Investigate the finding
Follow the decision trace and inspect a reduced reproduction of a policy issue.
Investigate issues - 04
Examine what changes
Compare two explicit rule sets over shared journeys in the What-if workspace.
Compare policies
A clearer view of policy
Understand what your policies actually do.
Start with a real example in the NovaCart workspace.