A working browser demonstration of a structured operational workflow. It is not presented as a deployed casino system.
Product maturity
Interactive workflow demonstrationKnow exactly what this page represents.
A workflow-fit review, customization scope, implementation plan, and a decision on whether the workflow should remain a browser tool or become a controlled production application.
Diagnostic analysis workflow
Surveillance Risk-Signal Analysis
Review surveillance and operational risk signals using verified counts, baselines, trends, severity, evidence preservation, procedure gaps, alternative explanations, confidence, access controls, escalation, ownership, and deadlines.
Workflow demonstrationAnalytics workspaceReady for workflow-fit reviewHow demonstrations are controlled →
Six evidence-led analysis modules
Compare all analysis modules →Casino Operational Analytics Workbench
Current analysis module: Surveillance Risk-Signal Analysis · Risk-signal analysis
Diagnostic case
Determine whether rising surveillance signals represent more risk or better detection
Monthly surveillance counts increased after camera upgrades, a revised referral procedure, and additional analyst coverage. The raw increase could reflect greater exposure, improved detection, duplicate referrals, changed classification, or an actual control problem.
Incident Dashboard shows a recurring or rising signal category, but management needs a multi-period analysis before changing controls, staffing, procedures, coverage, or investigative priorities.
After exposure, detection opportunity, taxonomy changes, evidence quality, duplicates, and false positives are controlled, which recurring conditions justify a proportionate operational response?
A restricted analytical pack with controlled signal definitions, rates and severity, evidence confidence, detection and coverage context, procedure-gap mapping, competing explanations, escalation thresholds, and a bounded control test approved by authorized management.
Surveillance Risk-Signal Analysis isolates one specific operating decision
This page is built around the exact failure, evidence standard, approval boundary, and implementation conditions that make Surveillance Risk-Signal Analysis different from the other workflows in the library.
Where the number becomes misleading
Incident Dashboard shows a recurring or rising signal category, but management needs a multi-period analysis before changing controls, staffing, procedures, coverage, or investigative priorities.
Why one total or percentage cannot answer the question
Incident Summary documents one event, Surveillance Incident Dashboard controls the active case portfolio, Review Request Intake authorizes a bounded review, and Game Protection Review evaluates a specific sequence. This analytics workflow examines many controlled signals across time. It controls taxonomy, exposure, detection opportunity, duplicates, evidence confidence, false positives, procedure conditions, and coverage changes so management can test a control weakness without ranking people or declaring misconduct from counts, proximity, or correlation.
The exact management judgment supported by the analysis
After exposure, detection opportunity, taxonomy changes, evidence quality, duplicates, and false positives are controlled, which recurring conditions justify a proportionate operational response?
A restricted analytical pack with controlled signal definitions, rates and severity, evidence confidence, detection and coverage context, procedure-gap mapping, competing explanations, escalation thresholds, and a bounded control test approved by authorized management.The records and definitions that must reconcile first
- Signal taxonomy and version
- Defines the exact signal category, inclusion and exclusion rules, classification version, severity scale, confidence level, and authorized use of the measure.
- Exposure denominator
- Captures relevant table hours, transactions, camera-hours, access events, reviews, guest contacts, or other approved opportunity used to calculate a meaningful rate.
- Detection environment
- Records camera coverage, image quality, system availability, analyst staffing, referral volume, rule changes, training, and other factors that alter detection opportunity.
- Evidence and disposition
- Separates verified events, unresolved signals, duplicates, benign explanations, false positives, evidence holds, retention status, and final authorized disposition.
What must be standardized before managers compare results
- Approve a versioned signal taxonomy with inclusion, exclusion, duplicate, severity, confidence, disposition, privacy, and retention rules before trend analysis.
- Reconcile referrals, incidents, evidence holds, final dispositions, camera and system availability, analyst coverage, procedure versions, and operational denominators.
- Separate verified, unresolved, duplicate, benign, false-positive, unsupported, and closed-with-limitation records in every count and rate.
Analytical purpose
The performance movement this workflow helps investigate
Analyze surveillance and operational risk signals using baselines, trends, severity, preserved evidence, procedure gaps, confidence, and escalation.
Review surveillance and operational risk signals using verified counts, baselines, trends, severity, evidence preservation, procedure gaps, alternative explanations, confidence, access controls, escalation, ownership, and deadlines.
Analytical inputs
Define the measure before explaining the movement
The same number can mean different things when definitions, denominators, time windows, or operating conditions change.
- 01
Approved operational datasets
- 02
Metric definitions and comparison period
- 03
Known data gaps, assumptions, and source references
- 04
Chronology, attributable observations, evidence status, and access restrictions
- 05
Defined metrics, comparison basis, source totals, and material variance notes
Diagnostic record
What the analysis must capture beyond the headline metric
These fields separate a plausible driver from a convenient story and keep uncertainty visible during review.
Signal taxonomy and version
Defines the exact signal category, inclusion and exclusion rules, classification version, severity scale, confidence level, and authorized use of the measure.
Exposure denominator
Captures relevant table hours, transactions, camera-hours, access events, reviews, guest contacts, or other approved opportunity used to calculate a meaningful rate.
Detection environment
Records camera coverage, image quality, system availability, analyst staffing, referral volume, rule changes, training, and other factors that alter detection opportunity.
Evidence and disposition
Separates verified events, unresolved signals, duplicates, benign explanations, false positives, evidence holds, retention status, and final authorized disposition.
Condition and control map
Links recurring signals to procedure steps, physical zones, technology states, staffing conditions, handovers, temporary controls, and known control gaps.
Escalation and test record
Defines rate or severity thresholds, restricted recipients, proportional safeguards, test duration, guardrails, review authority, and conditions for ending escalation.
A rise in chip-transfer alerts is separated from improved camera coverage
- Recorded chip-transfer signals doubled over three months after two overhead cameras were replaced and supervisors began submitting a new standardized referral form.
- Total transfer volume also increased, and several referrals describe the same event from both pit and cage perspectives.
- Most reviewed events are procedurally correct, while a smaller subset shows delayed acknowledgement during shift change rather than evidence of theft or collusion.
- The source set includes referral records, incident dispositions, camera uptime, transfer volume, procedure versions, staffing, evidence holds, and review timestamps.
The analysis deduplicates paired referrals, calculates verified and unresolved rates per thousand transfers, separates pre-upgrade and post-upgrade detection environments, grades evidence confidence, and maps the remaining recurring condition to delayed shift-change acknowledgement. It recommends a limited acknowledgement-control test rather than an allegation-focused response.
The Surveillance Manager, Cage Manager, Table Games Manager, and Compliance approve the restricted taxonomy, denominators, evidence treatment, privacy boundaries, temporary safeguard, test duration, escalation rule, and wording. Individual misconduct conclusions remain outside the analytics workflow and require separate case evidence and authority.
A management-ready output—not just a completed form
The working app organizes the result so management can understand the position, verify the evidence, choose an action, record approval, and assign the next review without rewriting the workflow from scratch.
What the completed workflow should make clear
Analysis prepared from approved inputs, with source references, open questions, named ownership, limitations, and a visible management review point.
Analysis
The decision management must make
After exposure, detection opportunity, taxonomy changes, evidence quality, duplicates, and false positives are controlled, which recurring conditions justify a proportionate operational response?
The app prepares the decision; it does not approve or execute it.Records that should support the recommendation
- Approved operational datasets
- Metric definitions and comparison period
- Known data gaps, assumptions, and source references
- Chronology, attributable observations, evidence status, and access restrictions
What management still needs to question
- Comparing signal counts across periods after camera, staffing, referral, training, or classification changes without adjusting for altered detection opportunity.
- Using incidents, referrals, alerts, observations, allegations, duplicates, and verified events as interchangeable units in one trend line.
- Calculating rates without a relevant exposure denominator such as transfers, table hours, access events, transactions, reviews, or camera-hours.
Responsible department head or General Manager
This reviewer confirms the decision record. The complete approval gate is stated once in Operational boundaries.
Close the action with ownership and a checkpoint
Prepared by: Authorized analyst or reporting coordinator · Responsible department record owner
Next checkpoint: The reviewer sets the follow-up date, confirms the responsible person, and records whether the matter is closed, monitored, returned for correction, or escalated.
What should remain after the meeting
- Operating position
- The analysis deduplicates paired referrals, calculates verified and unresolved rates per thousand transfers, separates pre-upgrade and post-upgrade detection environments, grades evidence confidence, and maps the remaining recurring condition to delayed shift-change acknowledgement. It recommends a limited acknowledgement-control test rather than an allegation-focused response.
- Decision owner
- Responsible department head or General Manager
- Status
- Draft, reviewed, approved, returned for correction, monitored, or closed
- Required record
- Evidence references, approved action, responsible person, approval status, follow-up date, and remaining uncertainty
Who prepares the comparison set
- Authorized analyst or reporting coordinator
- Responsible department record owner
The preparer should preserve definitions, time windows, comparison groups, known confounders, and any missing observations.
Who validates the interpretation
Responsible department head or General Manager
Final approval requirements are consolidated in the Operational boundaries section below.
What management must decide for this workflow
Only the controls that are specific to this application are shown here. The shared portfolio standard is documented once in the methodology.
Approved data, accountable review, management authority, and evidence-based claims apply across the portfolio.
How demonstrations are controlled →- Responsible reviewer
- Responsible department head or General Manager
- Decision before use
- Responsible department head or General Manager approves the prepared analysis and assigns any follow-up before it is shared or used.
- Not for
- Do not use this to identify individuals, determine guilt, or bypass restricted access, evidence preservation, and escalation rules.
- Application-specific limits
- It does not determine intent, guilt, disciplinary action, or a final incident conclusion.
6 workflow-specific risks to review
These are practical failure risks for this workflow, not repeated portfolio-wide disclaimers.
- Comparing signal counts across periods after camera, staffing, referral, training, or classification changes without adjusting for altered detection opportunity.
- Using incidents, referrals, alerts, observations, allegations, duplicates, and verified events as interchangeable units in one trend line.
- Calculating rates without a relevant exposure denominator such as transfers, table hours, access events, transactions, reviews, or camera-hours.
- Treating repeated proximity, shared shift, nationality, guest status, employee identity, or another protected or contextual attribute as a risk conclusion.
- Allowing restricted evidence, identities, investigative details, retention deadlines, or unresolved allegations to appear in broad management outputs.
- Deploying a safeguard that shifts activity into blind areas, disrupts operations, overwhelms analysts, or creates more false positives than useful evidence.
Lock definitions and comparisons before the pilot
- Approve a versioned signal taxonomy with inclusion, exclusion, duplicate, severity, confidence, disposition, privacy, and retention rules before trend analysis.
- Reconcile referrals, incidents, evidence holds, final dispositions, camera and system availability, analyst coverage, procedure versions, and operational denominators.
- Separate verified, unresolved, duplicate, benign, false-positive, unsupported, and closed-with-limitation records in every count and rate.
- Model changes in detection opportunity caused by equipment, coverage, staffing, training, referral behavior, classification, and review backlog.
- Map recurring conditions to procedure steps, zones, system states, handovers, workload, temporary controls, and evidence for competing explanations.
- Design one proportionate control test with restricted access, operational and privacy guardrails, analyst capacity limits, escalation criteria, and exit conditions.
How to test whether the workflow improves explanations
- A second authorized analyst can reproduce signal definitions, deduplication, dispositions, denominators, rates, severity, confidence, and coverage adjustments.
- The workflow visibly changes interpretation when camera uptime, referral practice, classification, or analyst coverage changes detection opportunity.
- Verified events, unresolved signals, duplicates, benign explanations, and false positives remain separate in every summary and management statement.
- No person, team, guest group, or protected attribute is ranked or labeled suspicious solely through association, proximity, frequency, or correlation.
- The selected safeguard produces measurable control evidence without unacceptable operational disruption, privacy impact, analyst overload, or false-positive growth.
- Authorized management records continue, modify, expand, withdraw, or escalate decisions with evidence, limitations, restricted distribution, and review date.
Test the diagnostic logic
Challenge the assumptions before using the conclusion.
Analyze surveillance and operational risk signals using baselines, trends, severity, preserved evidence, procedure gaps, confidence, and escalation.