Local operational decision support
OpsProof proves the safest operational change—or refuses to invent one.
Use a compatible local Fulfillment or Support event log to find where pressure forms, test a disruption, and identify the lowest-cost staffing change that qualifies in the modeled scenario.
CSV only · stays in your browser · no API key · scenario planning, not a guarantee. Check exact CSV requirements.
- 0198%Normal SLA reliability
Evidence check
Compatible Fulfillment event log
Local timestamps, activities, and resources map to one reviewed operation.
- 0233%Stress SLA reliability
Modeled disruption
Packing station outage
The same operation is tested with a demand shock and one unavailable station.
- 0395%Recovery reliability
Bounded recovery
Add 1 packing capacity unit
For two hours—the lowest-cost tested option that qualifies in the modeled scenario.
One evidence chain, three comparable states. Scenario planning only; no recommendation appears unless the tested evidence qualifies.
Before you upload
Will my event log work?
OpsProof checks this exact contract locally before it calibrates a model. It does not repair missing or incompatible operational evidence.
- File: UTF-8 CSV, up to 1 MB and 10,000 event rows.
- Required columns: Case ID, activity, start time, completion time, and resource.
- Time: ISO-8601 timestamps with
Zor an explicit UTC offset. - Evidence: at least 20 completed cases across 60 observed minutes.
- Supported operations: reviewed Fulfillment or Support workflows only.
You will define the service-level target (SLA): the percentage of completed cases that must finish within a stated time. CSV templates and compatible/rejected examples are available in the next step.
Product walkthrough
Watch the 76-second product demo.
See the compatible input, modeled disruption, recovery comparison, and the evidence that keeps every recommendation bounded.
The demo uses reviewed, synthetic operational examples and makes no claim of real-world predictive certainty.
What you receive
A decision you can inspect, not a black-box prediction.
- 1. Measure pressure. Identify the replicated bottleneck across the same fixed-seed runs.
- 2. Test disruption. Stress the confirmed operation with visible assumptions.
- 3. Compare recovery. Show the lowest-cost capacity option that meets the stated qualification threshold.
One bounded change restores the tested operation.
- Normal service reliability
- 98%
- Under modeled stress
- 33%
- Tested recovery
- Add 1 capacity unit to Packing stations for 2 hours · 95% qualifying reliability · $56
- Mean missed-target orders
- 24.92 → 2.93
Stress → Recovery across matching 100 runs · 21.99 prevented per 480 simulation-minute operating horizon
Packing had the largest queue in 96 of 100 runs. Seed fingerprint: seeds-v1:6b05a6ec. This is scenario-planning evidence from fixed-seed simulations, not a real-world guarantee.
Ready to check your data?