Proposal · May 2026
Intelligence
for Maurya
Global
AI agents built specifically for your factory floors — capturing data that's currently being lost, alerting on problems before they become rejections, and giving every department an expert that never sleeps.
Agents in this proposal
01
Production
Production Intelligence
Real-time machine tracking, idle alerts, tooling ID
›
02
Quality Assurance
QA Intelligence
Cp/Cpk analysis, computer vision defect detection
›
03
Foundry
Foundry Intelligence
Batch monitoring, composition alerts, pattern learning
›
Production runs
on paper and
memory today
Paper-based tracking — production orders tracked in hard copy, no live visibility for management or supervisors
No idle detection — machines sit idle for hours with no alert raised; utilisation rates are unknown
Tooling by memory — wrong tool on a precision part is discovered only after the damage is done
Current visibility
0%
of machine utilisation is measured digitally
All tracking done manually on paper at end of shift
No live production signal
Alerting
None
Real-time alerts
Machine idle → no notification sent
Tracking
∅
Live order tracking
Status lives in supervisor's head
QA runs
entirely
on paper
Measurement data on paper — dimensional readings written in sheets that cannot be compared, searched, or analysed
No rejection tracking — rejected parts logged informally with no batch, shift, or reason recorded consistently
Visual inspection by eye — inspector on the 200th part of a shift will miss cracks a camera would catch every time
Data capture
100%
of QA measurement data recorded on paper
No digital record survives past end of shift
No trend line, no history
Capability
0
Cp/Cpk analysis
Process capability never calculated
Gate check
None
Pre-production checks
No material verification before a run
A bad pour wastes
an entire heat
of 1,200 kg
No real-time composition alerts — spectro results are checked, but deviations are not flagged until after the pour proceeds
Sand property inconsistency — readings taken manually and inconsistently, no pattern tracking across shifts
No batch traceability — when QA rejects a casting, there is no structured way to trace it back to the foundry conditions that produced it
Batch risk
1,200 kg
wasted per heat when chemical composition is wrong
4 furnaces — one bad batch cascades across the full pour cycle
High-cost error surface
Prevention
None
Pre-pour alerts
Deviation caught after, not before
Root cause
∅
Batch traceability
Rejection ↔ foundry cause: unknown
Not a generic
ERP. Built for
Maurya Global.
✓
Configured for your factory — your four furnaces, your alloys, your part drawings, your rejection patterns
✓
Intelligence that compounds — Month 1: captures lost data. Month 3: surfaces patterns. Month 6: actively reducing rejections
✓
One dashboard for everything — Foundry, QA, Production live on one screen without walking the factory floor
Deployment Timeline
Month 1
Production + QA agents deployed. Data capture begins. Rejection tracking goes digital for the first time.
Month 2
Foundry agent live. Camera-based visual inspection operational at QA station. Per-heat logs accumulating.
Month 3
Pattern learning surfaces first insights. Rejection rate trends visible across departments. Unified dashboard live.
Month 6+
The system knows Maurya Global's factory better than any off-the-shelf tool ever could. Intelligence compounding every batch.