Try engines with sample data
Every engine ships with a pre-loaded example scenario — a realistic construction-industry use case you can run in one click, no documents required.
Why sample data
New users often stall at the "what do I upload?" step. Sample data lets you skip that — click Try with sample data on any engine card, and the engine renders a fully-populated result for a realistic scenario. You see the output, the callouts, the exports, and the chain suggestions before committing to your own data.
Sample runs are rendered from canned output — they don't call the AI and don't consume your runs quota. They're identical to what a real analysis produces on the same inputs, frozen at a point in time. Use them to orient yourself before spending a real run.
The sample scenario
We standardize all sample data around one fictional project: Bridgeport Datacenter Phase 2 — a 180-MW hyperscale build in southwestern Connecticut, $45M electrical subcontract, 8 bidders, 26 weeks from RFQ to substantial completion.
Using one scenario across all engines means you can trace the same vendor (Apex Electric), the same contract, the same change orders through every engine. You see how engines connect.
| Engine | What the sample shows |
|---|---|
| RFQ Generator | Generated $45M electrical RFQ with 212 line items, exclusions, inclusions. |
| Bid Leveling | 7 qualifying bids normalized side by side; Apex recommended; one unbalanced bid flagged. |
| Scope Compliance | One vendor's bid vs the RFQ — two silent exclusions found ($340K + seismic). |
| Contract Review | Apex's AIA A401 subcontract with 41 clauses, 7 flagged (LD rate too high, broad pay-if-paid). |
| Contract Summary | 2-page executive brief of the same contract for the PM. |
| Change Order Review | $847K CO #3 analyzed — 62% outside original scope, recommended payment $548K. |
| Pay App Review | Pay App #1 audit: 3 red flags, retainage math error, $2.31M recommended vs $2.8M requested. |
How to use it
Open any engine
Go to Engines in the app and click on the engine you want to explore.
Look for the 'Try with sample data' chip
Most engines show a small chip near the top of the input panel. If it's there, click it.
See the rendered output
The engine skips straight to results, populated with Bridgeport data. Every section — summary, tables, risk flags, chain suggestions — is real output that the engine would produce on real inputs.
Dismiss when ready
Click Start fresh in the results header (or navigate away) to clear the sample and return to your real workflow.
Chain buttons on sample results are visible so you can see what they do, but clicking one starts a real engine on empty inputs. That's intentional — the sample is for orientation, not a pre-built demo project.
Engines without samples yet
The tracker engines (COI Tracker, RFI Tracker, Owner Report, Field Daily Report, Punch List, Submittal Tracker, Procurement Tracker) don't have sample scenarios yet — they produce tracker rows rather than standalone reports, and sample-izing them would require seeding fake tracker state. We'll add them once that pattern settles.