Every record. Every dollar. Every decision.
A law firm's billing file enters on the left. Nothing becomes payable until a rule or a certified human says so. This demonstration walks that path in six steps — about three minutes.
- Press Run. Six billing cycles are reviewed deterministically in the browser.
- Follow the money. The bar below never moves without a stated cause.
- Take the steps in order. Findings → decisions → QC → export.
The payable amount changes only when authority exists.
Run the review to establish the controlled populations.
Run the review and watch where records go
What this shows: the six-stage path every billing record travels. What to do: press Run deterministic review above, then watch which records exit to Cleared and which are held for a human.
Original intake → controlled downstream output
- 01LEDES intakeBytes and lineage preserved
- 02Deterministic reviewStructure, arithmetic, OCG controls
- 03Consensus if requiredEvidence only; never approval
- 04Human reviewIndependent disposition and certification
- 05Cleared for processingAuthorized population
- 06Downstream outputVerified, resumable, duplicate-safe
CONTEXTWhat population is being reviewed?
Acme, Pinky, Brain & Associates
A deterministic, reproducible six-cycle synthetic firm. Every person, matter, hour, dollar, defect, and disposition is inspectable in later steps.
One row, one cause, one dollar effect
What this shows: the governing control, the reason, and the review path for each record. What to do: filter to Human Review to see only the contested records, or select any matter link to open that record in full.
PPR control view
A dollar never moves without a stated cause.
| Matter / invoice | Line | Original | Dollar change | Result | Rule / control | Why passed or failed | Review involvement | Final disposition |
|---|
Select a matter to open the full record in Step 3.
Original → DRE review → resulting invoice
The same records, shown three ways. Compare the tabs to see exactly what changed.
The machine record, translated without changing it
What this shows: a single billing record in plain language, with the underlying JSON one click away. What to do: step through records, then select Inspect underlying JSON to confirm the readable view hides nothing.
Record detail
Nothing clears without a named, certified human
What this shows: the two controlled populations and the reviewer workspace. What to do: select a held record, enter any reviewer name, choose a disposition, tick the certification box, and watch the payable figure above move.
Human review and cleared for processing
Reviewer evidence, QC evidence, and export history persist in this browser until reset.
Human review
Cleared for processing
Measure deterministic quality without signalling the sample
What this shows: concealed sampling of the cleared population and the resulting defect rate. What to do: set a rate, lock the sample, then open it — the selected records are not identified anywhere until you do.
Locked sample, independent verdict
Sampling draws only from the deterministically cleared population.
One record at a time, verified before the next
What this shows: a controlled downstream adapter with duplicate prevention and recovery state. What to do: export cleared records, then open the reviewed CSV and resulting LEDES file to check the output yourself.
Verified output
A resulting LEDES 1998B invoice becomes available only when every source line for that invoice is cleared and exported.
- Batch identity
- —
- Adapter
- Canonical reviewed CSV + LEDES 1998B
- Recovery state
- CLEAN
- Verified / queued
- 0 / 0
- Duplicate prevention
- ACTIVE
- No downstream attempts recorded.
Aggregate measurement, economics, and history
Not part of the six-step walkthrough. These surfaces roll the same record population up by operating unit, billing cycle, and labour cost. Select any row to drill back to its underlying records.
DRE Sheet
Plan, actual, variance, financial movement, throughput, quality, and labour, derived from the current record population. Select an operating unit to drill into its records.
| Operating unit | Plan hours | Actual hours | Variance | Original billing | Adjustments | Resulting billing | Records | Cleared | Human review | Throughput | QC sample | Defects | DPMO | Review labour |
|---|
Illustrative firm model; not customer operating data. Negative movements appear in parentheses.
Good billing moves; judgment stays controlled.
QC labour calculator
1,000 records × 7 minutes × $25/hour
Illustrative billing funnel
Benchmark assumptions supplied for demonstration context.Firm and client billing cycles
| Cycle | Client | Matters | Hours | Submitted billing | Adjustments | Resulting billing | Records | Defects | DPMO |
|---|
Three questions. Under a minute.
Your answers open a pre-filled email in your own mail client — nothing is transmitted from this page. Candid criticism is more useful than encouragement.
No file upload is accepted and no matter data leaves the browser. Consensus results are simulated evidence, not approval. The controls shown are illustrative and must be configured, validated, and authorized for each production environment.
Official LEDES source file