AI for water utilities — with NIS2 evidence

Assistant for network documentation and emergency procedures

Knowledge and documents

Problem: Network documentation and emergency procedures are scattered across departments

How it works: A central RAG knowledge base with OCR of diagrams and citations of the right procedure

What you gain: The duty officer finds the right emergency procedure and diagram with a citation in seconds — also at night, without calling a colleague

Assistant for letters and correspondence with residents

Process automation

Problem: Handling residents’ applications and correspondence overloads the customer office

How it works: The assistant prepares draft replies based on the case documentation; integration with e-Doręczenia (e-Delivery) and ePUAP

What you gain: The customer office replies faster and more consistently, and staff approve a draft instead of writing from scratch

AI maintenance with limited IT resources

Integration and reliability

Problem: A small IT team lacks the resources to maintain an AI system after deployment

How it works: CogOps (CDF-F6) — a monthly ongoing service with a dedicated consultant and answer quality monitoring; fallback models with automatic failover

What you gain: The AI system has a permanent caretaker and a contingency plan — the IT team does not have to become model specialists

Automatic summaries of operational data

Analytics and reporting

Problem: Meter readings and operational data require regular summaries

How it works: Schedules (RRULE) and in-chat code execution for calculations on operational data

What you gain: Operational summaries are generated regularly from repeatable calculations — to check, not to assemble

Secure and controlled use of AI

Security and sovereignty

Problem: Residents’ and customers’ data must not reach an external provider

How it works: Personal data masking; an “EU only” switch; a predefined GOV / Energy sector profile

What you gain: Customer data is masked automatically, and the sector policy is ready from the start — no need to build it from scratch

Evidence package and AI oversight for an operator of an essential service

Compliance and oversight

Problem: As an operator of an essential service, you must document oversight over the use of AI

How it works: AI agent register (CDF-F2); the CDF Platform generating a unified evidence package for audits; an immutable event log

What you gain: Documentation of AI oversight for KSC/NIS2 is produced as the system works — it is not a project before the audit

Are our systems (SCADA, leak detection) high-risk systems?

AI that serves as a safety component in the management of water supply is listed in Annex III to the AI Act. The Initial Assessment in the CDF methodology settles this for each use case

Does the AI tool meet the NIS2 supply chain requirements?

SAIE provides the documentation needed to assess the supplier: data location, a processing log, the manufacturer’s ISO/IEC 27001 and 42001 certificates and an exit plan

How do you link the AI policy to the information security management system?

The AI policy becomes part of the ISMS: the same roles, asset register and review. Templates for the policy and registers are in the WDR documents