IRIS — Intelligent Analytics Platform

AI Assistant

LUMA

The intelligence layer of IRIS

IRIS sees.
Luma illuminates.

Ask questions in Arabic or English. Get instant answers grounded in your real data — no SQL, no BI training, no waiting.

lumen — Latin for light لُمعة — Arabic for gleam AR + EN Bilingual Voice Input Grounded · No Hallucination

"IRIS is the eye that sees all your data. Luma is the light that makes it visible. Without light, the eye cannot function."

✨

Luma

Online · Connected to your data

What were the top 3 performing regions last quarter?

Based on Q3 data: (1) Kuwait City — 34% of revenue, ↑8% QoQ. (2) Riyadh — 28%, ↑12% QoQ. (3) Dubai — 19%, ↑5% QoQ. Shall I break this down by product category?

ما هي المؤشرات التي تجاوزت الهدف هذا الشهر؟

تجاوزت 4 مؤشرات الهدف هذا الشهر: معدل الاحتفاظ بالعملاء (94%)، صافي نقاط الترويج (72)، وقت الاستجابة (8 دقائق)، ومعدل إتمام المكالمات (87%).

🔍SQL from natural language

📊Auto chart + narration

🎙️Voice input (Custom ASR)

📚Knowledge base RAG

🔒Tenant-scoped only

🤖Agent builder

LiveBoards

Not a dashboard.
A LiveBoard.

A LiveBoard is a living, communicating workspace — not a static report. Every widget is connected to a live data channel with configurable thresholds, AI narrative, and multi-channel alerts.

Traditional Dashboard

Static. Passive. Read-only.

✗Built by IT, consumed by users
✗Fixed refresh schedule
✗No personalization
✗No proactive alerts
✗No AI explanation

IRIS LiveBoard

Live. Communicating. Yours.

✓Built by any user — no IT
✓Real-time via data channels
✓Personalized per role
✓Email · WhatsApp · Teams alerts
✓Luma AI narrative per widget

📊 KFF Emergency Readiness — Kuwait LIVE

Active Centers

of 42 total

Avg Readiness

77%

↑ 3% this week

Open Incidents

2 critical

✨ Luma Narrative

Readiness improved 3% this week driven by Farwaniya and Ahmadi centers completing equipment maintenance. 2 critical incidents in Hawalli — nearest qualified center is 4.2 km away with 89% readiness score.

Alert Channels

📧 Email 💬 WhatsApp 🔔 In-App 📱 SMS 🟦 Teams

Platform Features

Everything you need.
Nothing you don't.

108 production-grade features across 15 functional categories — from data ingestion to AI-powered decision support.

🧠

Luma AI Assistant

Natural language analytics in Arabic and English. Generates SQL, charts, and narrative explanations automatically from your real data.

Example

A telecom analyst asks "Which IVR nodes have the highest drop-off?" — Luma queries, builds a funnel chart, and explains which nodes need redesign.

📋

LiveBoards

User-built real-time monitoring workspaces. Each widget connects to a live data channel with thresholds, alert rules, and AI narrative.

Example

A fraud team builds a LiveBoard with 8 widgets — when any metric exceeds threshold, WhatsApp alert fires within 30 seconds.

🏗️

KPI Semantic Layer

Visual KPI designer with formula builder, dimension hierarchies, target tracking, and multi-step approval workflow. One definition, consistent everywhere.

Example

A municipality defines "Asset Utilization Rate" once — formula, dimensions, targets, and owner. Every dashboard and AI answer uses the same definition.

🚨

Early Warning System

Multi-condition alert rules with AND/OR logic, Isolation Forest anomaly detection, escalation chains, and alert analytics. Proactive — not reactive.

Example

Alert fires when fraud rate >3% AND volume >10,000 in any 1-hour window — escalates to fraud manager and auto-creates an investigation ticket.

✅

Data Quality Index

6-dimension automated quality scoring: completeness, accuracy, timeliness, consistency, uniqueness, and validity. Auto-recomputes with drill-down to field level.

Example

12% of citizen records have missing phone numbers. DQI flags this, Luma explains the impact on SMS delivery, and a remediation task is auto-created.

🗺️

GIS & Geospatial AI

Google Maps integration with KPI heatmaps, choropleth layers, clustering, and AI-powered incident response scoring. Click the map to trigger spatial analytics.

Example

KFF: click any location on Kuwait map → enter incident type → IRIS scores all 42 centers by skills, equipment, distance, staff → recommends top 3 responders.

📈

Forecasting & Anomaly Detection

ARIMA/SARIMA time-series forecasting with MLflow experiment tracking. Isolation Forest anomaly detection with explainability and confidence intervals.

Example

A mining company forecasts phosphate production for 6 months. Model confidence: 94%. Anomaly detected: production dip in Week 23 — traced to maintenance schedule.

🛡️

Fraud Detection & AML

ML-based fraud scoring with velocity checks, merchant risk profiling, network analysis, and AML compliance monitoring. Case management with investigation workflow.

Example

Merchant with 340% velocity spike in 2 hours. Fraud score: 94/100. System auto-flags, creates a case, notifies analyst, and suggests blocking pending review.

🎙️

ASR Voice Bot Analytics

Arabic + English speech recognition analytics. Track WER/CER, intent recognition rates, language-switching patterns, fallback/escalation analysis, and call outcome intelligence.

Example

A bank's Arabic voice bot has 78% intent accuracy. IRIS identifies the top 12 misunderstood intents and recommends training data additions to reach 92%.

🛒

FMCG & Retail Intelligence

Sales analytics by SKU/region/channel, distribution coverage, promotions ROI, shelf intelligence, supply chain KPIs, and demand forecasting.

Example

FMCG distributor tracks 1,200 SKUs across 8 regions. IRIS identifies 3 SKUs causing 67% of stockouts and forecasts a 23% demand spike in Ramadan.

🔌

Multi-Source Connectors

Connect DuckDB, PostgreSQL, MySQL, SQL Server, Oracle, REST APIs, Excel/CSV, and cloud warehouses. Incremental ELT with lineage tracking and schema versioning.

Example

Telecom connects Oracle billing + PostgreSQL CRM + REST API for network KPIs + Excel for budget targets — all unified in one semantic layer.

🤖

Agent Builder

Build custom AI agents with defined tools, scopes, and guardrails. Agents query data, trigger alerts, generate reports, and interact with external systems — all audited.

Example

A "Daily Risk Briefing" agent runs every morning at 7am: queries overnight transactions, identifies top 5 risk patterns, generates a PDF, and emails the risk committee.

Industry Verticals

Built for your industry.

IRIS is not a generic BI tool. Each vertical receives a purpose-built analytics experience — domain-specific KPIs, workflows, and AI models.

Banking, Fintech & Payment Processing

Real-time fraud detection, AML compliance monitoring, transaction analytics, and risk intelligence for banks, payment processors, and fintech companies across GCC and MENA.

✓
ML-based fraud scoring with velocity checks and merchant risk profiling
✓
AML transaction monitoring with suspicious activity reporting
✓
Card portfolio analytics — approval rates, decline reasons, chargeback trends
✓
Merchant performance and risk segmentation
✓
Real-time EWS alerts with case management workflow

Sample KPIs

Fraud Rate2.4%▼ Target <3%

AML Alerts / Day47▲ +12%

Chargeback Rate0.8%▼ -0.2%

Approval Rate94.2%▲ +1.1%

Avg Fraud Score31/100▼ Low risk

Telecom — IVR, Roaming & ASR Analytics

End-to-end analytics for telecom operators — from IVR call journey intelligence to roaming revenue analytics (RoamIQ) and ASR voice bot performance monitoring.

✓
IVR call journey analysis — node-by-node drop-off and completion rates
✓
ASR accuracy tracking — WER/CER by language, dialect, and intent
✓
Roaming revenue analytics by partner, country, and service type
✓
Subscriber segment analysis — language preference, call behavior
✓
Network generation performance (2G/3G/4G/5G) by region

Sample KPIs

IVR Self-Service Rate68%▲ Target 75%

ASR Intent Accuracy84%▲ +6% MoM

Agent Transfer Rate22%▼ -4%

Roaming RevenueKWD 2.4M▲ +18% QoQ

Avg Handle Time3m 42s▼ -18s

Government & Public Sector

KPI governance, PMO portfolio management, citizen service analytics, and GIS-integrated spatial intelligence for government agencies, municipalities, and programme offices.

✓
PMO portfolio — project health, milestone tracking, budget vs actual
✓
Citizen service analytics — response times, satisfaction, complaint resolution
✓
GIS-based KPI heatmaps by district, governorate, or zone
✓
National KPI dashboards with multi-level approval governance
✓
Budget execution monitoring with variance alerts

Sample KPIs

Projects On-Track73%▲ +5%

Budget Utilization68%▼ Target 80%

Citizen Satisfaction4.2/5▲ +0.3

Avg Response Time2.4 days▼ -0.8d

KPI Compliance Rate91%▲ +3%

Emergency Services — KFF Fire Force

Real-time operational readiness monitoring, AI-driven incident response recommendations, staff skills management, and equipment availability tracking across all centers.

✓
Incident Response Recommender — AI scores all centers for any incident type
✓
Center readiness scoring — skills, equipment, staff availability, proximity
✓
Staff skills gap analysis — certified vs required per incident type
✓
Equipment maintenance and operational status tracking
✓
Annual performance reports and training compliance

Sample KPIs

Avg Center Readiness77%▲ +3%

Certified Staff Rate84%▲ +2%

Equipment Availability91%▼ -1%

Response Time (Avg)6.2 min▼ -0.8m

Incident Resolution Rate98.4%▲ +0.4%

FMCG & Retail

Unified view of sales performance, distribution coverage, promotions ROI, and demand forecasting — across all channels, regions, and SKUs — for FMCG manufacturers, distributors, and retailers.

✓
SKU-level sales analytics by region, channel, and time period
✓
Distribution coverage and stockout detection by area
✓
Promotions ROI analysis — uplift vs baseline, cannibalization detection
✓
Demand forecasting with seasonal adjustment (Ramadan, holidays)
✓
Retailer performance ranking and incentive tracking

Sample KPIs

Distribution Coverage87%▲ +3%

Stockout Rate4.2%▼ -1.1%

Promo ROI+34%▲ +8%

Forecast Accuracy91%▲ +4%

Revenue per SKUKWD 4,820▲ +12%

Asset Management — Municipalities & Utilities

Asset lifecycle analytics, Remaining Useful Life (RUL) predictions, financial exposure modeling, and maintenance optimization — integrating with Maximo and other CMMS systems.

✓
Asset portfolio overview — condition, age, category, department
✓
RUL prediction — which assets will fail in the next 12 months
✓
Financial exposure — replacement cost by risk tier
✓
Maintenance budget planning with scenario modeling
✓
Data quality monitoring for asset registry completeness

Sample KPIs

Assets at High Risk12%▼ -2%

Avg Asset Age8.4 yrs▲ +0.3

Replacement ExposureAED 42M▼ -8%

Maintenance Compliance88%▲ +5%

DQI Score91/100▲ +3

Mining Operations

Production analytics, financial performance monitoring, and sustainability tracking — with forecasting models for production planning and cost optimization.

✓
Production volume analytics by mine, shift, and product type
✓
Revenue and cost per tonne with variance analysis
✓
Production forecasting (ARIMA + seasonal adjustment)
✓
Sustainability KPIs — energy, water, emissions per tonne
✓
Equipment utilization and downtime impact analysis

Sample KPIs

Production Volume184K t▲ +7%

Cost per Tonne$28.4▼ -$1.2

Equipment Uptime94%▲ +2%

Forecast Accuracy93%▲ +1%

CO₂ per Tonne0.42 t▼ -0.03

Demo Videos

See IRIS in action.

Watch IRIS demonstrate real analytics scenarios across different industries. Demo videos are being produced.

Watch Demo

IRIS Platform Overview

Complete walkthrough — data connection to LiveBoards, Luma AI, and vertical modules.

⏱ ~5 minutes

Coming Soon

Luma AI — Ask your data anything

Luma AI in Action

Watch Luma answer complex analytics questions in Arabic and English in real time.

⏱ ~3 minutes

Coming Soon

LiveBoards — Build & Alert

Building a LiveBoard

A business user builds a LiveBoard in minutes — widgets, alert channels, and sharing.

⏱ ~4 minutes

Coming Soon

Fraud Detection — Banking Demo

Fraud Detection Demo

Real-time fraud scoring, merchant risk profiling, and AML alert workflow.

⏱ ~6 minutes

Coming Soon

KFF Incident Response — Map Demo

KFF Incident Response AI

Click on Kuwait map, enter incident type, watch IRIS score all centers and recommend the optimal response team.

⏱ ~4 minutes

Coming Soon

ASR Voice Bot Analytics

Arabic + English call bot performance — intent accuracy, dialect confusion, escalation analysis.

⏱ ~5 minutes

Feature Showcase

See IRIS in the real world.

Real screens. Real data. Real decisions — from LiveBoards to root-cause causal trees, these are actual IRIS features in production.

LiveBoard + Slice by Dimension

Bulletin Board — AI Collaboration

Bulletin Board — Intelligence Feed

KPI Relationship Designer — Canvas

KPI Relationship Designer — List

Root Cause Analysis — Tree

Root Cause Analysis — List

LiveBoard

LiveBoards

Slice any dimension.
Instantly.

This is a live MOE Basra Operational Account board. The user clicked a dimension — City/Governorate — and the entire board filtered in real time. Baghdad: 553K transactions. Karbala: 191K. Every widget on the board responded simultaneously.

Dimension types

0ms

Filter propagation

∞

Widgets per board

Dimension Registry Board-wide filter Sticky notes Export PDF Timeline Report Share

Not a dashboard — a LiveBoard

Bulletin Board

Where data meets
conversation.

A team member asks: "which merchant in Baghdad has the highest return rate?" — @alsaqi (Luma AI) responds inline with a chart, causal narrative, and a drill-down link. No context switching. No separate BI tool. The insight lives in the conversation.

@alsaqi

Mention AI in any channel

Charts to LiveBoard

AI inline response Drill-down from chat Export PNG Pin to Board @mention team Channel threads

Analytics inside the conversation

Bulletin Board

Intelligence Feed

Bulletin Board — Feed

Your personalised
intelligence feed.

Every morning, IRIS delivers a Morning Brief — top 3 KPI movements, anomalies, and AI insights — directly to your feed. Live KPI ticker scrolls across the top. Pinned KPIs show real-time values. Alsaqi Radio broadcasts the brief in Arabic audio.

1 min

Auto-refresh interval

Audio broadcast

LIVE

KPI ticker

Morning Brief Alsaqi Radio Pinned KPIs AI Insight cards Weather widget Live ticker

Intelligence, not information overload

KPI Relationship Designer

Map every KPI
relationship.

19 KPIs. 116 relationships. Mapped, validated, and approved. The canvas shows every causal connection — Direct, Inverse, Threshold, Lagged — with edge thickness representing causal path weight. This is the semantic backbone that powers root-cause analysis.

116

Relationships mapped

Approved

KPI categories

Causal graph Auto-Layout Direct / Inverse Threshold / Lagged Strong / Moderate / Weak Correlations view

The semantic layer behind every insight

IRIS KPI Relationship Designer canvas view

KPI Designer — Canvas

KPI Designer — List

KPI Relationship Designer

Every KPI.
Fully governed.

The list view shows every KPI with its formula, polarity, live value, trend, direction, owner, approval status, and relationship count. The Actions menu exposes Root-Cause Analysis, Slice by Dimension, Generate Audio Report, and Pin to Bulletin Board — all from a single row.

711K

Live transaction count

+7.3%

Trend this period

Approval workflow Root-Cause from row Audio Report Pin to Board Health Score Why Role (Leaf/Pillar)

Governed KPIs, not spreadsheet metrics

Root Cause Analysis

Why did this
KPI move?

AML Approval Rate moved. IRIS traces the causal chain — upstream driver: AML Average Transaction Value (+1.3%, causal path weight 9.0%). Downstream effects: Transaction Volume and AML Decline Rate (-6.8%). One click: Ask IRIS to explain, Create EWS Alert from the top driver, or Open a Diagnostic Session.

9.0%

Causal path weight

+1.3%

Top driver movement

Causal tree AI Summary Ask IRIS to explain Create EWS Alert Diagnostic Session Download SVG

From symptom to root cause in seconds

Root Cause — Tree

Root Cause — List

Root Cause Analysis

Ranked drivers.
Quantified impact.

The list view ranks every upstream driver by causal path weight with trend sparklines for the last 12 periods. Downstream effects (pillars) are listed with their current values. Every driver is tagged Strong or Direct — so analysts know exactly where to focus.

Driver ranked by causal weight

Periods of trend history

Causal path weight Trend sparkline Strong / Direct tags Downstream pillars Ask IRIS Create EWS Alert

Quantified causality, not guesswork

AI-Powered Insights. Real-Time Decisions. Across Every Vertical.

IRIS sees.
Luma illuminates.

Not a dashboard.
A LiveBoard.

Everything you need.
Nothing you don't.

Connect any data source.