Top critical issues — composite criticality score TWIN
Ranked by composite criticality — click a row for Asset 360°, sort any column, search or filter
All
Crit
High
Med
| Asset | Zone | Customers | Score | Severity |
|---|
Resumen General — service by region DERIVED · Cognos
Full-utility rollup · 1.33M client accounts · 112 filtration plants
| Region | Plants | In service | % service |
|---|
Avoided-cost model TWIN
Benefit of acting = reduction in outage + emergency-response + non-revenue-water cost vs. deferral
Deferral exposure (open orders)—
Avoided 5-yr cost (portfolio)—
Outage cost run-rate—
Non-revenue-water (modeled)$1.9M / yr
Blended benefit-cost ratio—
Budget vs. value ($M) FY2026
Capital + operating resilience program
committed —headroom — / $24.5M
| Line | Cost | Avoided 5-yr |
|---|
Predicted failures — next 30 days watsonx.ai
Trend-extrapolated from the operational twin — cumulative lifeline-asset failures, surfaced ahead of service loss
| Asset | Health | Pred. fail | Cause | Age open |
|---|
Generator readiness gap EGU
713 power facilities · readiness drives outage exposure in grid events
92.6%
Operable
53
Not operable
31
Pending repair
22
Pending replace
OperablePending repairPending replace
Investment & expenditure priorities — ranked by BCR TWIN
Prioritized portfolio · benefit-cost ratio = avoided cost ÷ remediation · fund from remaining FY budget
Check items to fund — committed, headroom & blended BCR recompute live; click a row for Asset 360°
| Asset | Customers | Total cost | Avoided 5-yr | BCR | Est. payback |
|---|
FY2026 budget execution
committed YTD —FY budget headroom —
Selected portfolio · items—
Total cost—
CAPEX / OPEX— / —
Blended BCR—
Parts & inventory — 30-day projection Materials
Maintenance-parts demand & reorder exposure (mission c)
30-day parts cost$146k
Lines below reorder point5
Stockout risk (≤2 on hand)3
| Critical part | On-hand | Status |
|---|
DRRS Operations Assistant RAG
Grounded answers over governed PRASA sources · incident corpus, procedures, SAP-PM work orders
⛨ Grounded (RAG)No external egressPII maskedAudit logged
Capabilities & connected sources
Capabilities
- Grounded (RAG) answers over the governed PRASA corpus
- Composite-criticality ranking & downstream impact
- Drafts budget requests & SAP-PM work orders (human-approved)
Connected sources
Grounded in PRASA dataYes
External egressNone
PII handlingMasked
Audit logCaptured
Governed by watsonx.governance · on-premises · no write-back except the human-approved SAP work order via watsonx.orchestrate.
Purpose
This briefing accompanies the PRASA Use Case Request to IBM. It demonstrates, in navigable form, the executive results of replacing the incumbent Foundry-based DRRS on an IBM read-only analytics stack — extended with failure prediction and parts/inventory-cost projection.
Modules demonstrated
- Executive Command — Resumen General regional rollup & composite-criticality situation
- Financial & decision — avoided-cost, BCR, FY budget execution
- Asset health & predicted failure — watsonx.ai, ahead of service loss (absent in the incumbent)
- Generator readiness & parts/inventory projection
- DRRS Operations Assistant — grounded, governed RAG
Evidence basis
Operational findings derive from the PRASA incident corpus (~96,650 records) and DRRS workspace exports; the Resumen General is the derived (Cognos) report layer. The operational KPIs here are computed live on a 39-node San-Juan twin sample that mirrors the production ontology.
Verification path
- Incumbent contract values — PR Office of the Comptroller contract registry
- Budget figures — PRASA approved FY2026 capital & operating budget and execution records
- Costs — Independent Cost Estimate (ICE) and the IBM priced response
Framework alignment
Centralized real-time twin · predictive maintenance · geospatial + multi-signal fusion · human-validated repair orders prioritized by projected cost-of-deferral · AI for failure prediction.
Integrity notice. This is a consultant-produced interactive mock-up illustrating the requested
target state of the IBM solution; it is not a production system. Operational values are from a representative twin sample.
Design-stage · structure mirrors the PRASA DRRS study mock-up, re-skinned in the project design system. Operational KPIs from the 39-node twin sample; Resumen General / generators / parts from the study datasets (derived report layer).