Skip to main content
This section describes the end-to-end operational flow of Likemind Safe.

8.1 Audit Workflow

STEP 1: SETUP (Owner, one-time)
├── Configure Factory Locations (Setup → Locations)
├── Configure Hazard Categories (Setup → Hazard Categories)
└── Configure Risk Matrix (Setup → Risk Matrix)

STEP 2: CREATE AUDIT SESSION (Owner or Auditor)
├── Set title, location, date
└── 1 session credit is consumed
    └── Credit priority: Rollover → Plan → Top-up

STEP 3: ADD CHECKPOINTS
├── Manual entry: title, location, description
└── AI-Suggested: system generates based on context

STEP 4: COLLECT EVIDENCE
├── Upload before photos per checkpoint
└── Annotate photos if needed (canvas editor)

STEP 5: AI ANALYSIS (per checkpoint)
├── System analyzes photos + context + knowledge base
├── Returns: hazard category, risk level, findings, recommendations
├── Citations rendered as [Ref: ...] and [Per: ...] badges
└── Auditor reviews and adjusts AI output

STEP 6: FINALIZE
└── Session is complete when all checkpoints are reviewed

8.2 CAP Workflow

STEP 1: CREATE CAP SESSION (Owner or Auditor)
└── Group related findings under one session

STEP 2: ADD FINDINGS
├── Link findings to audit data (hazard category, risk level)
├── Assign PIC, deadline, follow-up date
└── Status starts as "pending" or "in_progress"

STEP 3: EVIDENCE COLLECTION
├── Upload after photos showing remediation
└── Annotate photos if needed

STEP 4: SUBMIT FOR REVIEW
├── Single session: Submit for Review button
├── Bulk: Submit multiple sessions at once
└── Status changes: Draft → Submitted

STEP 5: REVIEWER EVALUATES
├── Reviewer opens Reviewer Portal or external link
├── Reviews findings, photos, recommendations
├── Adds review comments
└── Decision: Approved OR Revision Required

STEP 6: ITERATION (if Revision Required)
├── Auditor addresses reviewer feedback
├── Updates findings and evidence
└── Resubmits → back to Step 5

STEP 7: COMPLETION (if Approved)
├── CAP marked as approved
└── Approved findings auto-extracted to Knowledge Base
    └── Future AI analyses reference these learnings

8.3 Session Billing Mechanism

CREDIT TYPES
├── Plan Credits: Allocated per billing period by subscription tier
├── Rollover Credits: Unused plan credits from previous period (50% cap)
└── Top-up Credits: Purchased one-time packs (never expire)

CONSUMPTION ORDER (when creating an audit session)
1. Rollover credits consumed first
2. Plan credits consumed second
3. Top-up credits consumed last

MONTHLY RESET
├── Unused plan credits × 50% = rollover for next period
├── Top-up credits are NOT reset (they persist indefinitely)
└── Sessions used counter resets to 0

PLAN CHANGE
├── 100% of remaining credits roll over (not 50%)
└── New plan credits are allocated immediately

WARNING SYSTEM
├── Warning banner appears when nearing session limit
└── Hard limit dialog blocks session creation when quota is exhausted

8.4 Knowledge Feedback Loop

Approved CAP → Auto-extracted to Knowledge Base

Knowledge Base (CAP Knowledge + Uploaded Documents)

AI Checkpoint Analysis references Knowledge Base

Better, context-aware findings and recommendations

New CAP findings...
This creates a continuous improvement cycle: each resolved CAP makes future AI analysis more accurate and relevant to your specific organization.