AIMM Index
User Manual v0.8 · IS360 Technologies & Services Pvt. Ltd.
Make in India MAKE IN
INDIA
OVERVIEW

Introduction to AIMM Index

What is AIMM Index?

AIMM Index is a continuous compliance and governance twinning platform. Instead of producing a governance report once per quarter, AIMM Index recomputes each organisation's AI Maturity Score (AI$M) on every telemetry event — translating your AI model's real-time behaviour into an auditable maturity level from L0 to L6.

🇮🇳 Made in India
Developed and operated by IS360 Technologies & Services Pvt. Ltd., Chennai, India, under the Make in India initiative.

AIMM Level Scale (L0–L6)

LevelLabelAI$M RangeWhat it means
L0Ad Hoc0–19%No consistent AI governance
L1Aware20–34%Awareness exists, processes informal
L2Repeatable35–49%Some repeatable controls
L3Managed50–64%Measured and tracked
L4Optimised65–74%Proactive continuous improvement
L5Advanced75–84%Systematic, data-driven governance
L6Production Ready85–100%Fully governed, audit-ready
ℹ️ AI$M Formula
AI$M = 100 × √(ACI × SIR) — ACI (Accuracy/Consistency Index) from confusion-matrix telemetry; SIR (System Integrity & Resilience) from encryption, integration, modularity, and continuity scores.
ACCESS

Login & Account Access

Signing In

1
Open the platform
Navigate to app.aimmindex.com. The public landing page appears first. Click Launch Platform or Sign In.
2
Enter credentials
Email and password as provisioned by your administrator.
3
Session token issued
An 8-hour signed session token is stored in your browser and sent automatically with every API call.
⚠️ No self-service signup
Accounts are provisioned manually. Register at aimmindex.com/#register or email partha@is360technologies.com.

Role Capabilities

RoleRegister twinsTelemetryCalibrateAnnexureAdmin pageChat
Viewer
Operator
Auditor
Admin

Two Separate Frontend Files

URLPurpose
app.aimmindex.com/ (index.html)Public landing page — product info, guide, pricing, registration
app.aimmindex.com/dashboard.htmlRBAC command-centre — live KPIs, Growth Comparison, Annexure, Certification
app.aimmindex.com/console.htmlOperational console — register twins, telemetry, calibrate, agent chat
TWIN MANAGEMENT

Compliance Twins

What is a Compliance Twin?

A compliance twin is a persistent digital governance record for one AI-enabled organisation or system. Each twin holds: its confusion matrix (TP/TN/FP/FN), system integrity parameters, a calibration baseline, the full AI$M time-series (aism_history), and the active domain agent's latest findings.

Registering a Twin

ADMIN ONLY
1
Open the Operational Console
console.html → Register Twin in the sidebar.
2
Fill required fields
org_id (unique, no spaces), name, domain (determines which agent monitors this twin — see Domain Agents section).
3
Submit
Twin is created at L0, uncalibrated.
ℹ️ Domain field is permanent
The domain value selects the monitoring agent for this twin's lifetime. Choose carefully: "Semiconductor Fab", "HealthTech", "Retail MSME", "Logistics", "Automotive", "EV", "EduTech", "Agri" — see the keyword map in the Domain Agents section.

Calibrating a Twin

AUDITOR+

Calibration locks the current AI$M as the baseline. Drift alerts only fire after calibration.

1
Submit at least 5 telemetry events
Cover TP, TN, FP, FN, drift, and at least one SIR-layer score.
2
Review current AI$M
Confirm the score is genuine before locking it.
3
Calibrate
Console → Lifecycle → Calibrate, or POST /api/twin/:orgId/calibrate.

Twin States

Statetwin_typeMeaning
UncalibratedstaticRegistered, may have telemetry, no baseline set. Drift not tracked.
Calibrated/LiveliveBaseline locked. Drift tracked. Annexure shows full history.
TELEMETRY

Submitting Telemetry

Telemetry Metric Reference

metric_typeMeasuresComponentRange
TPTrue positivesACIInteger ≥0
TNTrue negativesACIInteger ≥0
FPFalse positives (false alarms)ACIInteger ≥0
FNFalse negatives (missed detections — highest stakes in HealthTech)ACIInteger ≥0
driftProcess/signal drift — each agent interprets differentlyACIFloat 0–1
encryption_scoreCryptographic postureSIR: EsFloat 0–1
integration_scoreAPI/system integration completenessSIR: EiFloat 0–1
modularity_scoreArchitectural independenceSIR: MpFloat 0–1
continuity_scoreUptime / resilience — tightest floor for HealthTechSIR: CrFloat 0–1
L0–L6Layer-specific governance score overrideLayer scoresFloat 0–100
⚠️ Minimum viable set
You need at minimum: TP, TN, FP, FN, drift, and one SIR-layer score to produce a non-zero AI$M.

Submitting via the API

POST /api/twin/telemetry
Authorization: Bearer <token>

{"org_id":"your-org","metric_type":"TP","value":780}

The response includes updated aism, aimm_level, drift object, and agent_alerts array — findings the active agent raised from this event.

DOMAIN AGENTS

Domain-Specialized Agents

How Agent Selection Works

Domain keyword(s)Agent selectedDrift alert threshold
semiconductor, fab, chipSemiconductor Fab8% (tighter)
retail, d2c, ecommerceRetail / MSME12%
logistics, fleet, delivery, freightLogistics10%
automotive, auto, ev, electricAutomotive & EV10%
health, medtech, clinical, diagnosticHealthTech6% (tightest)
edu, learning, lmsEduTech16%
agri, farm, cropAgri20% (loosest)
(anything else)Generic fallback15%
ℹ️ One formula, many specialized eyes
The AI$M formula never changes. What changes per domain is which drift percentage is dangerous, which telemetry imbalances matter most, and what uptime floor is acceptable.

HealthTech Agent — Tightest Thresholds

HIGH CONSEQUENCE — TIGHTEST THRESHOLDS
CheckThresholdCode raisedSeverity
Diagnostic miss-rate (FN/total)>4%HEALTH_MISS_RATEwarning/critical
Monitoring drift≥3%HEALTH_MONITORING_DRIFT_WARNINGwarning
Monitoring drift≥6%HEALTH_MONITORING_DRIFT_ALERTcritical
Continuity/uptime (Cr)<0.7HEALTH_UPTIME_RISKcritical
⚠️ FN is weighted specially
False negatives (missed detections) are weighted as the dominant risk in HealthTech — a missed diagnosis is worse than a false alarm. The agent monitors FN/total specifically, not FP+FN together.

Semiconductor Fab Agent

CheckThresholdCodeSeverity
Yield variance (FP+FN/total)>5%FAB_YIELD_VARIANCEwarning/critical
Particle drift≥4%FAB_PARTICLE_DRIFT_WARNINGwarning
Particle drift≥8%FAB_PARTICLE_DRIFT_ALERTcritical
Uptime (Cr)<0.6FAB_UPTIME_RISKwarning

Why tighter than generic: fab contamination compounds quickly. A 9% drift that the generic agent ignores triggers a critical alert here.

Retail / MSME Agent

CheckThresholdCodeSeverity
Directional forecast bias (|FN–FP|/total)>15%RETAIL_FORECAST_SKEWwarning/critical
Return/fraud drift≥6%RETAIL_FRAUD_RATE_WARNINGwarning
Return/fraud drift≥12%RETAIL_FRAUD_RATE_ALERTcritical
Inventory continuity (Cr)<0.5RETAIL_TURNOVER_RISKwarning
ℹ️ Skew direction matters
Positive skew (FN > FP) = stockout risk (under-forecasting). Negative skew = overstock/dead-stock risk. The agent calls out direction-specific recommendations.

Logistics Agent

CheckThresholdCodeSeverity
ETA prediction bias (|FN–FP|/total)>15%LOGISTICS_ETA_SKEWwarning/critical
Cargo damage/loss drift≥5%LOGISTICS_DAMAGE_RATE_WARNINGwarning
Cargo damage/loss drift≥10%LOGISTICS_DAMAGE_RATE_ALERTcritical
Fleet utilisation (Cr)<0.5LOGISTICS_FLEET_UTILISATION_RISKwarning

Automotive & EV Agent

CheckThresholdCodeSeverity
Defect/rework rate (FP+FN/total)>6%AUTO_DEFECT_RATEwarning/critical
Process drift (non-EV)≥5–10%AUTO_PROCESS_DRIFT_WARNING/ALERTwarning/critical
Battery/thermal drift (EV)≥5–10%AUTO_EV_BATTERY_DRIFT_WARNING/ALERTwarning/critical
Line/charging uptime (Cr)<0.55AUTO_UPTIME_RISKwarning

EV context is auto-detected from the domain field — battery/thermal language replaces generic process language when "ev" or "electric" appears.

EduTech Agent

CheckThresholdCodeSeverity
At-risk learner miss-rate (FN/total)>8%EDU_DROPOUT_MISS_RATEwarning/critical
Assessment integrity drift≥8–16%EDU_INTEGRITY_DRIFT_WARNING/ALERTwarning/critical
Platform continuity (Cr)<0.45EDU_PLATFORM_CONTINUITY_RISKwarning

Agri Agent — Sparse Data Aware

CheckThresholdCodeSeverity
Low telemetry volume<5 eventsAGRI_SPARSE_TELEMETRYinfo
Yield prediction skew (|FN–FP|/total)>20%AGRI_YIELD_SKEWwarning
Crop-stress drift≥10%AGRI_CROP_STRESS_DRIFT_WARNINGinfo
Crop-stress drift≥20%AGRI_CROP_STRESS_DRIFT_ALERTwarning
ℹ️ AGRI_SPARSE_TELEMETRY is informational
Agri telemetry is seasonal — gaps between harvest and sowing are normal. The agent flags sparse data so you are aware, not to escalate it.

Generic Agent — Fallback

CheckThresholdCodeSeverity
AI$M drift from baseline≥8%DRIFT_WARNINGwarning
AI$M drift from baseline≥15%DRIFT_ALERTcritical
Uncalibrated with ≥5 events≥5 eventsCALIBRATION_DUEinfo
Any layer score below floor<35/100WEAK_LAYERwarning
No telemetry for 7 days7 daysSTALE_TWINwarning
DASHBOARD

Command-Centre Dashboard

Overview Page — What is Live vs Reference

PanelDataStatus
AI$M KPI / Twin count / AlertsGET /api/twin/list + agent endpointsLive — auto-refresh 20s
Growth Comparison chartGET /api/twin/compare (real aism_history)Live — real data
AI/Agent Insights feedAgent alerts from all twinsLive
Annexure download per twinGET /api/twin/:id/annexure (generated live)Live — real .docx
Global Distribution MapIllustrative marker positionsStatic placeholder
Certification StatusStatic reference data (flagged in-app)Manual
Training & CompetencyRole-scoped static reference dataManual
Audit TrailStatic reference data (flagged in-app)Manual

Growth Comparison (Geometric)

VIEWER+
1
Navigate
Dashboard sidebar → Growth Comparison.
2
Select twins
Tick at least 2 twins.
3
Plot
Chart renders from real aism_history snapshots. Calibration events are marked with larger circles.
ℹ️ Submit telemetry first
Growth history accumulates from telemetry events. Twins with no telemetry show "no history" in the legend.

Downloading a Compliance Annexure

VIEWER+

Click ⬇ Annexure next to any twin in the Compliance Twins page. Generates and downloads a live .docx containing:

  • Section 1: Twin Identity
  • Section 2: Current AI$M Compliance State
  • Section 3: Layer Score Breakdown
  • Section 4: Identified Gaps — findings from the active domain agent, run at generation time
  • Section 5: Growth History (last 15 snapshots)
  • Section 6: Attestation
ℹ️ Generated live at click time
Each Annexure reflects the twin state at that exact timestamp. Download a new one after each significant telemetry cycle.

AI Chat Agent

VIEWER+Requires ANTHROPIC_API_KEY on server
1
Select a twin
Confirm the active twin in the console.
2
Type your question
e.g. "Why did drift go to WARNING?", "Which layer is dragging my score down?", "What should I fix first?"
3
Receive role-scoped reply
Viewer: summary. Operator: remediation steps. Auditor: traceability. Admin: full detail.
⚠️ Read-only advisor
The chat agent explains scores and alerts — it cannot change any data. Requests to "set my score to X" are refused.
API REFERENCE

REST API Reference

Quick Reference

MethodEndpointMin RoleDescription
POST/api/twin/auth/loginAuthenticate. Returns signed token.
GET/healthHealth check.
POST/api/twin/registerAdminRegister new twin.
POST/api/twin/telemetryOperatorSubmit telemetry event.
GET/api/twin/listViewerList all twins.
GET/api/twin/:orgId/stateViewerFull twin state.
GET/api/twin/:orgId/driftViewerDrift analysis.
POST/api/twin/:orgId/calibrateAuditorLock baseline.
GET/api/twin/:orgId/agentViewerOpen alerts.
GET/api/twin/agent/sweepAuditorSweep all twins.
POST/api/twin/:orgId/agent/chatViewerChat with LLM agent.
GET/api/twin/compareViewerGrowth comparison (?orgs=a,b,c).
GET/api/twin/:orgId/annexureViewerDownload .docx Annexure.

Example — Register and Submit Telemetry

# 1. Login
curl -X POST https://api.aimmindex.com/api/twin/auth/login \
  -H "Content-Type: application/json" \
  -d '{"email":"partha@is360technologies.com","password":"YOUR_PASSWORD"}'

# 2. Register a twin (Admin)
curl -X POST https://api.aimmindex.com/api/twin/register \
  -H "Authorization: Bearer TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"org_id":"retail-demo","name":"Demo Retail Co","domain":"Retail MSME"}'

# 3. Submit telemetry (Operator)
curl -X POST https://api.aimmindex.com/api/twin/telemetry \
  -H "Authorization: Bearer TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"org_id":"retail-demo","metric_type":"TP","value":780}'
ANNEXURE & CERTIFICATION

Annexure & Certification

Free Certification — Promo Code AIMM-LAUNCH25

🎉 Launch Cohort Offer
First 25 organisations to register get one free AIMM Level Compliance Certification and Annexure run. Enter promo code AIMM-LAUNCH25 at aimmindex.com/#register. IS360 will contact you to provision access and walk through the initial calibration cycle.
1
Register at aimmindex.com
Enter promo code AIMM-LAUNCH25 in the registration form.
2
Await provisioning
IS360 provisions your account manually within the launch cohort.
3
Submit telemetry
Cover the minimum viable set: TP, TN, FP, FN, drift, one SIR-layer score.
4
Calibrate
Auditor or Admin locks the baseline.
5
Download your Annexure
⬇ Annexure link on the dashboard — your certified compliance evidence.
SECURITY

RBAC & Authentication Details

How Authentication Works

1
Password verification
PBKDF2 (100,000 iterations, SHA-256, 16-byte random salt per user). Never stored in plaintext.
2
Signed token issued
HMAC-SHA-256 over payload (email, role, expiry) using server-only AIMM_AUTH_SECRET. Only the server can verify.
3
Token sent with every call
Authorization: Bearer <token>. Server verifies signature before executing any handler.
4
Role enforced server-side
Even if a client forged a header claiming a higher role, the server rejects requests whose token does not carry the required role.
ℹ️ Dev-mode fallback
Without AIMM_AUTH_SECRET set, the API falls back to an unverified x-aimm-role header. This fallback is automatically disabled once the secret is configured — there is no way to accidentally leave it open in a properly configured deployment.

Adding New Users

ADMIN ONLY
// src/auth.js — add this line and redeploy:
seedUser("newuser@company.com", "operator", "New User", "NEW_USER_PASSWORD");

// Then set NEW_USER_PASSWORD in EB environment properties
⚠️ Redeploy required
Adding a user requires a backend redeploy. A user-management UI is on the roadmap, pending Postgres persistence.
PLATFORM STATUS

Known Limitations & Roadmap

What is Live vs Pending

FeatureStatus
AI$M engine + 7 domain agents✅ Live — 38/38 unit tests pass
RBAC / session auth✅ Live — real PBKDF2 + signed tokens
Annexure generation✅ Live — real .docx generated end-to-end
Growth comparison✅ Live — real aism_history time-series
Persistent data store (Postgres)❌ In-memory only — data lost on server restart
IoT / automatic telemetry ingestion❌ Manual API only — no MQTT bridge yet
Multi-user self-serve signup❌ Manual provisioning — requires backend redeploy
Live geolocation on map❌ Placeholder markers — no live coordinates per twin

Roadmap to TRL 7

TRL4→5
Persistent storage
Wire config/schema.sql to RDS Postgres. Data survives restart.
TRL5
Real pilot
2–3 of the 7 use cases with real organisational telemetry in their actual environment.
TRL5→6
IoT ingestion
MQTT/AWS IoT Core bridge for automatic telemetry from devices.
TRL6→7
Operational demonstration
Sustained multi-org operation with real governance reliance over a defined period.
AIMM Index User Manual · v0.8 · IS360 Technologies & Services Pvt. Ltd. · Chennai, India · © 2026 · partha@is360technologies.com