Funnel Conversion
—
Cases reaching each pipeline stage
Demo Mode — Synthetic data for evaluation only. Not real patient or case information.
Sign in
Request a Demo
Dashboard
Cases
Search evidence, reports, findings, and chronology within the selected case.
Demo preview
Live API connection is unavailable.
Business Associate Agreement
Version —
Draft agreement — not final legal text. This document is a placeholder pending legal review; it is shown exactly as stored, not rewritten.
Version · SHA-256 —
Send feedback
Help us make beta better. This goes to your organization's admins.
Authenticator setup
Add a time-based authenticator to your own account.
An authenticator is one control in your organization's HIPAA Security Rule program — it is not a certification. The secret and the backup codes are shown once and cannot be retrieved later.
Course
·
30 minutes · everyone (all ten organization roles)
What BluQube is for
It extracts clinical events from medical records, builds a chronology, identifies deviations from standard of care, and generates analysis reports from a unified platform.
Three things every beta user leaves with
- Who the users are. Ten organization roles: owner, admin, lnc_analyst, reviewer, client_viewer, attorney, paralegal, expert, nurse_reviewer, external_guest. The navigation shows about 22 surfaces; a given role sees a small subset — the server, not the UI, is the authority on what you may see.
- What a case looks like end to end. Records go in → 17 phases run → findings, a CRCM risk score, causal chains, and reports come out → approved reports become client deliverables.
- Nothing here is real. Beta runs on the synthetic case Doe v. Sunrise Nursing Home (2024-CV-1234). No real PHI is present. Beta is synthetic-data-only (board decision, 31 Oct).
The phase tracker, told honestly
Phase completion is derived from produced artifacts, not claimed. A fresh case's tracker starts at 0/17 and fills as phases produce artifacts. The seeded demo case shows 2/17 — the seeder writes only the chronology and deviation-analysis artifacts, so only phases 003 and 005 read complete. Trainers must never present "17/17 complete": that claim was the trust-killer finding UIX-B removed.
Client-side handout
Client viewers see four surfaces: Case Workspace progress, Case Timeline (read-only), Client Deliverables, and the Reports Center with approved items only. Download an approved deliverable; everything else is out of scope and out of the nav.
Canonical vocabulary and wording rules
- The canonical words are: Case, Evidence, Finding, AI Clinical Analysis, Review, Report, Deliverable. "Deviation" is a finding subtype; "error" and "mistake" are prohibited terms.
- Non-final work carries its visible marker:
DRAFT — NOT FINALorSeeded demo. Look for it; never remove it. - BluQube supports a customer's HIPAA Security Rule program — it is not a certification or attestation of one.
- Course completion confers nothing beyond itself: no licensure, no certification, no CE credit, no expert-witness qualification.
Mastery check (5 items, pass at 80%)
What the platform does; one canonical-vocabulary swap (Finding vs the prohibited terms); which roles are PHI-bearing; that beta data is synthetic; that the tracker fills from artifacts.
90 minutes + 60-minute hands-on lab (Module 2 rides inside this course) · analyst, reviewer, nurse_reviewer, admin, owner, attorney, paralegal
The 17 phases
Phase names and codes are verbatim from the phase manifest (CS-CLN-001…CS-CLN-017; shorthand below). Do not paraphrase phase names. Phases marked "no dedicated screen" or "pipeline-only" are facts about the current build per the owner's 31 Oct decision, not omissions.
| # | Phase | Where its output surfaces | Gate worth naming |
|---|---|---|---|
| 001 | Case Intake and Triage | Case Workspace #case (case focus, New Case quick action) | conflict check complete |
| 002 | Record Intake, Inventory, and Organization | Evidence Management #evidence (drop zone, inventory, SHA-256 integrity) | records cataloged, gaps classified |
| 003 | Chronology Development | AI Clinical Analysis #analysis (timeline); Case Timeline #timeline (client, read-only) | chronology sorted, source pages preserved |
| 004 | Standard of Care Analysis | No dedicated screen. The standards live in the case KB section standards_of_care (readable via GET /api/kb) and surface in three places: every deviation's detail panel cites the standard it is against ("Standard of care:" line); the AI summary reports a standards_identified count; and the dashboard coaching prompt "Standards of care not identified → Analyze Standards" appears until the section is populated. | hindsight-bias check complete |
| 005 | Deviation Identification | #analysis deviation markers; Case Workspace deviation summary; Review Queue #review | deviations link to a standard |
| 006 | Causation Analysis | ForenCoT View #forencot (causal chains, evidence gaps) | alternatives addressed |
| 007 | Expert Report Assembly | Reports Center #reports → Expert Report card / Report Studio | report sections complete |
| 008 | Oral Briefs and Attorney Consultations | No dedicated screen. The oral brief is a report kind (oral_brief) generated in Report Studio and listed in the Reports Center like any other report. It is LNC-internal attorney work product and is never promotable to FINAL — the promotion gate refuses it by kind (NEXUS-SEC-062). Its sections: key findings, open questions, gaps to raise, deviations to discuss, next steps. | counsel questions logged, follow-ups assigned |
| 009 | CRCM and ForenCoT Integration | CRCM Risk Map #heatmap + ForenCoT View | poster terms scored, deprecated framework not used |
| 010 | Deposition Preparation | Pipeline-only (owner decision 31 Oct): runs in the pipeline, state visible in the phase tracker; no dedicated screen. | vulnerabilities link to report opinions |
| 011 | Peer Review | Review Queue #review (approve/reject with required note, decision trail) | reviewer independence documented — see teaching rules |
| 012 | Discovery Response | Pipeline-only. | privilege review complete |
| 013 | EHR Analysis | Pipeline-only. | audit-trail source preserved |
| 014 | CME and Competency Tracker | Pipeline-only. | credentials current |
| 015 | Case Closure | Phase tracker; closure is a state, not a screen. | retention schedule set |
| 016 | Visual Aids and Exhibits | Reports Center (report cards); exhibits surface with their report/exhibit artifacts. | source data linked |
| 017 | Consolidated PDF Generation | Reports Center → Consolidated PDF / Export Package cards | all required phases complete |
Note: the user guide's "Phase Overview" table is a known defect (8 of 17, renames 002 and 009; fix ticket TRAIN-005). Do not teach from it — the table above is the manifest.
Teaching rules
- A phase flips to
completedonly when it produces a real artifact (UIX-B). The tracker is evidence, not a claim. - Gates are enforced by the pipeline (the manifest's completion gates). Skipping phases is the top cause of reports that do not survive a deposition.
- 011 and self-review (UIX-M). The owner decided (2026-11-01) that analysts may review their own case findings. The 011 gate is
reviewer_independence_documented— it requires the independence posture to be documented, and the platform does not and will not block an analyst from reviewing their own findings. Do not teach "the system prevents self-approval" — it does not. What training teaches instead: reviewers record who they are and why their review is adequate, and the decision trail shows it. - The Reports Center offers the report cards plus Report Studio (cited work products, provenance block). The "no dedicated screen" / "pipeline-only" rows above are stated as facts about the current build, per the owner's pipeline-only decision.
Module 2 — Hands-on lab: Doe v. Sunrise (60 min)
Setup: sign in at /login, open /?case_id=doe-v-sunrise#case. Work a scratch copy of the synthetic case so the seeded reference state (2/17 derived completions plus the seeded chronology/deviation artifacts) stays intact for demos — owner decision 31 Oct.
- Upload a record. Evidence Management
#evidence→ drop a file → it appears in the inventory with size and processing state. Reference set already on the case: admission_note_2023-06-01.txt, nursing_note_2023-07-01.txt, diet_record_july_2023.txt, discharge_summary_2023-08-01.txt, family_communication_log.txt. - Run the workflow. "Run Case Workflow" (all phases) or a single phase; watch the tracker fill as artifacts are produced; re-run safely (runs are idempotency-keyed).
- Read a deviation.
#analysis→ open a deviation and name four things: the standard cited, the severity, the confidence score, and the evidence file it links to. Demo anchors: dev-001 "no reassessment documented between June 5 and July 10 (35 days)" against "Braden Scale reassessment every 7 days" (moderate, 0.85); dev-002 "nursing notes indicate gaps of 4–6 hours" against "position changes every 2 hours for bedbound residents" (high, 0.90). Teach confidence as triage, not truth: HIGH ≥80%, MEDIUM 50–79%, LOW <50%. - Open the CRCM Risk Map
#heatmap. Demo values: Time 80, Delay in Care 75, Compliance 70, Documentation 65, Communication 60, Staffing 55, MedRisk 45, Skill Mix 40, Age 35, Legal History 15; 14 patterns matched (11 negative, 3 positive). Thresholds 0–20 Low … 81–100 Critical; the headline level floors at the worst factor — point at Time = 80 and show the headline cannot read below High. - Open ForenCoT
#forencot. Three causal chains; each link cites evidence. - Review a finding. Review Queue
#review→ open a pending item → approve or reject with a required note → confirm it appears in the decision trail. Note (UIX-M): an analyst may review their own findings; the trail records who reviewed and why. - Generate and download a report. Reports Center → Expert Report → open it; Export Package → download the ZIP. Confirm the approved report appears in Client Deliverables
#deliverablesas a client deliverable (expert_report.html, approved for client). - Recover from a failure. Read the failure message, retry the upload, use the offline banner's Retry; know the troubleshooting guide §2 (upload failures) and §3 (analysis timeouts).
Lab mastery is the checklist itself, signed off by observation (or by screenshot evidence for recorded/async delivery). A participant who cannot complete the upload and the review items repeats the lab; there is no partial lab credit.
Mastery check (pass at 80%)
Given a phase, name where its output surfaces (including "no dedicated screen — it appears in X"); state the artifact-derivation rule; name one completion gate; state what 011 requires after UIX-M.
COUNSEL APPROVED — final delivery decisions remain.
The source outline (EMPLOYEE-TRAINING-OUTLINE v1.1) remains "FOR REVIEW — NOT FINAL"; this beta-module approval does not change that source label. Delivery remains held on Security Officer curriculum sign-off and the Board's training-model decision.
Not yet deliverable to your role.
This module's content is available to internal workforce roles while final Security Officer and Board delivery decisions are pending. Client-side training remains checklist-tracked by the security officer.
45 minutes · PHI-bearing roles (owner, admin, lnc_analyst, reviewer, nurse_reviewer) — and attorney/paralegal per the register (delta D1; moves with the board answer via TRAIN-004)
What this module covers
- HIPAA fundamentals, condensed. Minimum-necessary is both a rule and the curriculum's own rule.
- Platform security, condensed. At-rest encryption is backups-only; session lifetime is 12 h default (the v1.1 corrections stand).
- Incident reporting — the one non-negotiable. Everyone must know how to report a suspected breach and to whom.
- Phishing awareness is optional here (beta participants are not LNC workforce); role-specific internal content lives in the role tracks.
Enforcement reality, stated plainly
- The PHI training gate is real and default-on. PHI-bearing role grants are refused without a current training completion; completions default to a 365-day validity window. Opting out requires an explicit setting and logs a startup warning.
- The gate is course-blind (delta D5): it keys on (person, role) and any current unexpired completion clears it —
LNC-ORIENTATIONalone included. The register's per-role course list is a curriculum contract, not the enforcement key. Never claim the gate checks which course was taken. - Client-side roles are not technically gated — they are customer personnel, not LNC workforce; their training is checklist-tracked by the security officer.
- Wording rule: the platform supports a customer's HIPAA Security Rule program; it is not a certification. No HIPAA claim ships in beta materials before counsel signs off.
Mastery check (pass at 80%)
Minimum-necessary applied to a role scenario; the report-a-suspected-incident route (what, to whom, how fast); what the gate does and does not check (including course-blindness); the approved HIPAA wording.
45 minutes · lnc_analyst, reviewer, nurse_reviewer
Scope of this track
Modules 1–2 plus a professional-standards block adapted from the LNT Factory Draft Syllabus v0.1 (Tom Stewart, 9 Oct — used with credit; the full program remains the Factory's, this is the beta slice). After it you can: upload evidence, run phases, read chronology/deviations/CRCM/ForenCoT, generate reports (analyst); triage the Review Queue, approve/reject with notes, read the decision trail (reviewer); clinical review of chronology, standards, and deviations with confidence-score triage and CRCM factor reading (nurse_reviewer). An analyst may review their own case findings (UIX-M); the decision trail records the reviewer.
Professional standards
- Scope and competence. Distinguish consulting work, clinical opinion, expert testimony, and legal advice; know when an assignment exceeds your specialty and how to refuse or refer; the professional retains responsibility for AI-assisted output — the platform's confidence scores are triage, never the reviewer's opinion.
- Records and evidence discipline. Every material finding carries a traceable source location; distinguish recorded fact, conflicting evidence, clinical inference, and unresolved question; absence of documentation does not by itself resolve what occurred — it is a gap to raise, not a conclusion.
- Work product. A defensible finding states its basis, its limitations, and its reviewer; separate clinical analysis from questions for counsel; communicate defects and corrections through the review trail.
- Critical failures (adopted list): wrong-patient/wrong-case material, unauthorized disclosure, fabricated or materially false citations, releasing an unverified material conclusion, or working beyond competence without referral. Any of these fails the track regardless of knowledge-check score.
Reference depth: the CRCM factors guide (cite it, do not summarize it).
Mastery
One deviation read aloud with all four attributes (standard, severity, confidence, evidence link) + one written scope statement naming what the reviewer will not opine on.
45 minutes · attorney, paralegal
Scope of this track
Read case progress and timeline, review reports and deliverables, download approved files. The workflow course teaches legal roles to read phase state, not run phases (delta D1). Paralegals add record-request tracking and evidence inventory checks, per granted scope.
What you are trained to judge
- Deliverables and reports. Only approved, final-labeled artifacts are admissible-looking.
DRAFT — NOT FINALmeans exactly that. - Oral briefs are internal work product and never FINAL — the promotion gate refuses the
oral_briefreport kind by design (see phase 008). - Privilege/discovery posture. Phase 012 runs a privilege review in the pipeline; there is no discovery screen in beta.
- Reading a deviation without over-reading an AI confidence score. Confidence is triage (HIGH ≥80%, MEDIUM 50–79%, LOW <50%), not a verdict.
Mastery
Given three artifacts (draft report, approved deliverable, oral brief), say which may leave the building and why.
45 minutes · owner, admin
Scope of this track
Operations surfaces — System Monitor, Security, Admin Panel, Organization Members, Audit Log — member role management, audit log reading, and incident reporting. Admins pause/resume AI and resolve operational exceptions. Owners do everything an admin can do, plus Billing & Subscription, organization governance, and training-exception approval.
What you are trained to operate
- Member role management, including that PHI-bearing role grants are refused without current training — the training gate (Module 4) — and that the security-officer exception path exists as a bound, expiring waiver, not a routine route.
- Audit log reading: membership changes, sign-in failures, and platform actions are recorded; training completions ride an audit event.
- Incident reporting: the route a suspected incident takes and who owns it.
Mastery
Add a member, find the audit entry, walk the incident-report route.
A completion is earned, not clicked: "Record completion" is accepted only after a knowledge-check attempt at or above the pass mark is on the server's attempt ledger. Practical items (the Module 2 hands-on lab and the Module 3 role-track observations) are recorded by an observer sign-off, which the knowledge check does not substitute for.
Course content is adapted from the beta training syllabus v1 (docs/training/BETA-TRAINING-v1.md); the course list and statuses come from the training register.
Workspace overview
Case operations at a glance
Monitor active cases, workflow progress, recent activity, and system readiness from one view.
Business Associate Agreement — acceptance required before case work
Document version — ·
A signed Business Associate Agreement is required before this organization can process PHI. Please ask your organization admin to accept the agreement from the dashboard.
✓ BAA signed — version —, signed on —.
View attestationHelp & resources
- Feature walkthrough
- Task map
- Best practices
- Troubleshooting
- Program overview
- API guides
- API reference
- Video tutorials — coming soon
Guides open in-app from the BluQube user-guide library — ask your administrator if a guide is missing.
Account security
Your sign-in authenticator for this account.
Loading authenticator status…
Analyst Workspace
Pipeline status, recent case activity, and analysis entry points for your organization.
Pipeline Status
Loading pipeline status…
Recent Case Activity
Loading case activity…
Analysis Entry Points
Active Cases
—
Pending Phases
—
across all active cases
Evidence Items
—
Quick Actions
Recent Activity
Across all cases in this organization
Loading recent activity…
System Health
AI Model Latency
—
Storage Used
—
API Status
—
AI Features
New in v1.11Recipe Search
Knowledge Graph
Knowledge Graph: Case Entities and Relationships
| Entity ID | Type | Description | Related Entities |
|---|
Medical Entity Extraction
Case: Doe v. Sunrise Nursing Home
About this page
- What it does
- The case's operational home: workflow progress, case focus, evidence inventory, and findings in one place.
- How it works
- Case work follows 17 dependency-aware SOP phases; the tracker reflects real phase state from the pipeline, which orders phases by declared dependencies and gates uploads before any phase runs.
- Why it matters
- Defensible work product starts with a defensible process — and the process is visible per case.
- What differentiates
- Phase order is enforced in code, not a kanban board someone can skip out of.
Delayed wound-care escalation
Unexplained delay in PRN pain-medication documentation
Case finding detail
Finding
Select a finding to review details.
AI Clinical Analysis
Review the case timeline, extracted entities, and deviations identified across the evidence.
Analysis results are case-scoped and loaded from the live API
About this page
- What it does
- Where the AI engines' candidate findings meet your judgment — deviations, extracted entities, and causal hypotheses for this case.
- How it works
- The chronology, entity-extraction, deviation-detection, and causal-review engines produce case-scoped results that this page loads from the live API. Every result is a candidate finding carrying source context and review status — nothing here is a final conclusion.
- Why it matters
- Machine reading finds the candidates; a qualified reviewer decides what they mean. This page is that handoff.
- What differentiates
- Findings arrive with their source context and review trail attached, not as a free-form summary you have to re-verify from scratch.
Clinical Timeline with Deviation Markers
03
03/12/2023 — Admission
Resident admitted with unspecified dementia, type 2 diabetes, coronary atherosclerosis, and osteoporosis. Braden Scale: 14 (moderate risk). Admission skin assessment records intact skin; the CMP calls for individualized repositioning and daily skin review.
04
04/15/2023 — First wound documentation
Nursing note: "blanching erythema, sacral area, 2×3 cm." Photo taken.
04
⚠ 04/22/2023 — Stage II pressure ulcer documented
Sacral wound now Stage II: "partial-thickness skin loss, 3×4 cm." Review point: No wound-care consult order appears in the sampled record by 05/17, although the facility CMP calls for escalation within 24 hours.
Confidence: 96%
|
Source: Progress note 04/22, MAR review
05
05/01/2023 — Braden reassessment
Score: 12 (high risk). 50 days since the admission score — compare the gap with the facility policy version in force.
05
⚠ 05/18/2023 — Wound care consult (26-day delay)
Wound-care clinician recommends off-loading, moisture management, dressing selection, nutrition screening, and follow-up. Review point: The consult occurred 26 days after the Stage II note; the complete treatment-order trail is not in the sample.
Confidence: 98%
|
Source: Wound consult note 05/18
06
⚠ 06/08/2023 — Fall incident
Resident found on the floor near the dining room; no injury was reported in the incident note. The timing of the post-fall assessment and care-plan review should be verified against the complete chart.
Confidence: 89%
|
Source: Incident report 06/08
08
⚠ 08/02/2023 — Wound progression to Stage IV
Sacral wound now Stage IV: "8×6 cm with 3 cm undermining." Review point: The record documents progression despite a wound-care plan; the available sample does not establish a single cause.
Confidence: 94%
|
Source: Wound assessment 08/02, MAR turning log
08
⚠ 08/15/2023 — Hospital transfer
Transferred to Memorial General Hospital for surgical debridement and IV antibiotics; the hospital record reflects a 12-day admission.
09
⚠ 09/15/2023 — Hospice referral
After hospital discharge and a goals-of-care discussion, the facility documented a hospice referral. The referral reason and any causal connection to the wound require independent review.
Medications
7 activeAcetaminophen 650mgTiming review
q4h PRN pain · 7-hour interval after a 7/10 score; request, reason, and reassessment are not in the sample
Furosemide 40mgas ordered
Daily · Last K+ 3.8 (normal)
Sertraline 50mgas ordered
Daily for depression/anxiety
Pantoprazole 40mgas ordered
Daily · GERD prophylaxis
Diagnoses
5 trackedUnspecified dementiaICD-10: F03.90
Documented at admission; severity and trajectory require chart review
Atherosclerotic heart diseaseI25.10
Documented at admission; no acute event asserted in this demo
Type 2 diabetes mellitusE11.9
Chronic condition documented at admission
Sacral pressure injury, stage 4L89.154
Stage sequence documented 04/22 → 08/02; acquisition and causation require review
Diagnosis Codes
read-onlyICD-10-CM codes documented in the record, with the source each code was read from. Display only — codes are not edited here, and a documented code is not a finding of causation.
Loading diagnosis codes…
Terminology (LOINC)
read-onlyDocument classifications and diagnosis assertions with review status. Display only — not edited here, and never a finding of causation or a CRCM score.
Document Types
Loading document types…
Diagnosis Assertions
Loading diagnosis assertions…
Expansion Packs
loadingInstalled expansion packs with version-pin integrity. Enabling or disabling a pack is platform-owner only and changes analysis behavior for every tenant. A listed pack is not a finding of causation.
Loading expansion packs…
Procedures / Interventions
4 notedWound care consult26d delayed
05/18/2023 · NP wound specialist
Wound-care plandocumented
05/18 · off-loading, moisture management, dressings, nutrition screening
Braden Scale assessmentsinfrequent
2 documented scores across 187 days; compare with facility policy
Hospice referralgoals-of-care
09/15/2023 · referral after hospital discharge and family discussion
About this page
- What it does
- The case's CRCM risk picture at a glance — factor scores, matched patterns, and how they cluster.
- How it works
- A rules engine scores clinical risk factors and matches CRCM patterns for the case; the heatmap shows the overall assessment, the worst single factor, and matched patterns, and a per-case co-occurrence graph shows how factors cluster. Portfolio-wide graphing is In Development.
- Why it matters
- It turns "something looks off" into named, reviewable pattern hypotheses a clinician can validate against the record.
- What differentiates
- The analysis follows a named clinical-compliance model — matches are candidate findings with review status, not conclusions.
BluQube / CRCM Intelligence
Forensic Risk Observatory
Case —
SYNTHETIC DEMO — NOT CASE EVIDENCE
CRCM Risk Index
—
—
Patterns
—candidate
recognized at selected time
Observed Protection
—
—
Data Status
—
—
Forensic Constellation
01 / BASELINE
Layers
Solid = documented link
Dashed = inferred association
Green solid = observed protection
Green dashed = hypothetical scenario
Psi amplification — Unavailable: model not established
Clinical chronology — plain-language reading
Same findings · simplified reading
Pattern matches are candidate findings supported by the cited evidence. Alternative explanations are not excluded; this view makes no liability or causation finding.
Definitions used in this view
- Standards of care
- Formal clinical standards — not "best practices" or informal guidelines.
- Deviation
- A documented departure from the standards of care. "Deviation" is the neutral, professional term — not "error" or "mistake".
- Causation
- The formal legal/medical link between a deviation and an injury. Links in this view show modeled risk reinforcement, not proven causation.
- Candidate pattern
- An Atlas pattern match supported by cited evidence — a finding to investigate, not a proven systemic failure.
- Hypothetical scenario
- An Atlas standard-of-care "what if" definition. Not a case finding and not an observed outcome.
Risk accumulation across episode of care
Illustrative
CRCM risk index
Event (observed)
Event (inferred)
Synthetic index values for layout demonstration, not clinical or forensic findings.
Observed protective signals are documented-presence detection only: a match shows a safeguard is documented in the case record — effectiveness is not assessed. When the matcher finds none, observed protection is reported as Not available. Hypothetical counter-patterns are Atlas standard-of-care definitions, not case findings.
Move the time cursor (or select Baseline, Escalation, or Cascade) to explore the illustrative progression. Values are mock data, not calculated CRCM outputs.
Single-run cases show the latest assessment as a point marker; the Observatory never interpolates a curve from one point.
Links depict modeled risk reinforcement and documented reasoning patterns. They do not independently establish causation; causation determinations remain analyst conclusions supported by the cited evidence chain.
Protective interventions show documented actions that could have interrupted this pathway in the model. They are illustrative counterfactuals, not proven clinical outcomes.
Organization-level delivery, reviewer load, and escalation queue — read-only.
About this page
- What it does
- Delivery status, reviewer load, and escalations across the whole organization.
- How it works
- Three live read surfaces — deliveries, reviewers, and escalations — assembled into one org-level queue view. Read-only by design: work is done in the case pages, observed here.
- Why it matters
- Review capacity is where legal-review businesses actually bottleneck; load has to be visible before it becomes a missed deadline.
- What differentiates
- Reviewer workload and escalations are first-class queues, not things you reconstruct from individual case views.
Deliveries
—Loading deliveries…
Reviewer Load
—Loading reviewer load…
Escalation Queue
—Loading escalations…
Organization-level KPIs — read-only. A KPI without a trusted data source shows why, instead of a number.
About this page
- What it does
- Organization-level operating KPIs in one read-only view.
- How it works
- Every value comes from the live KPI endpoint. A KPI whose data source is not trusted renders an explanation instead of a number — the page fails closed rather than guessing.
- Why it matters
- An operationally plausible fake number is worse than a visible gap — it gets repeated into decisions.
- What differentiates
- The fail-closed display is pinned by tests: values are live-sourced or honestly absent.
Turnaround Reliability
—
Accepted deliveries within 7 days
Citation Accuracy
—
Verified citations per analysis
Material Omission Rate
—
Analyses with a missed finding, of all CRCM analyses
Correction Burden
—
Deliveries sent back for revision
Escalation Rate
—
Escalations per active case
Reviewer Utilization
—
Active load vs. reviewer capacity
Collection Cycle
—
Days from invoice issued to paid
Client Concentration
—
Top-client share of revenue
Quality Incidents
—
Open incidents by severity
Case Timeline
Dated events from your case chronology — read-only, case-scoped data loaded from the live API
About this page
- What it does
- The dated chronology of your case — the spine every review is built on.
- How it works
- The chronology engine extracts dated events from extracted record text that keeps page-marker conventions, and the timeline is served per case. This view is read-only: corrections happen in the chronology workflow, not here.
- Why it matters
- Sequence is argument — gaps, delays, and ordering are often the case. Every event here can be traced to the page it came from.
- What differentiates
- Events are grounded in page-level source markers, not retyped summaries.
Loading timeline…
Keyboard Shortcuts
Press ? to toggle this help
Navigation
Go to Dashboard
Alt+1
Go to Case Workspace
Alt+2
Go to AI Analysis
Alt+3
Toggle Sidebar
Alt+S
Search
Open Search
Alt+K
Clear Search
Escape
Modals
Create New Case
Alt+N
Close Modal
Escape
Panels
Toggle Notifications
Alt+O
Toggle Dark Mode
Alt+D
Accessibility
Show Keyboard Shortcuts
?
Skip to Main Content
Tab (first press)
All shortcuts work with Alt key modifier