Overview

Sentinel is a platform built on graph database architecture and Reliability Centred Maintenance methodology. It gives engineers and technicians a conversational, intelligent interface to query complex asset data in real time — moving away from static spreadsheets and siloed records toward a living, queryable model of the asset base.

Currently in active development.

The Problem

Asset management data in most organisations is fragmented. Maintenance history lives in one system, asset registers in another, failure records somewhere else entirely. The people who need insight — engineers diagnosing a recurring fault, planners building a renewal case, managers assessing risk — spend more time hunting for data than using it.

Existing platforms are either too rigid, too expensive, or built for administrators rather than engineers. They answer the questions someone anticipated, not the questions that actually arise in the field.

The Approach

• Graph Database Architecture — assets, failures, maintenance tasks, components, and locations are stored as nodes and relationships, making it natural to traverse the connections that matter: what failed, what caused it, what was done, what is related.
• RCM Methodology — the data model is built around Reliability Centred Maintenance principles, ensuring that failure modes, consequences, and maintenance tasks are structured in a way that supports real engineering decisions.
• Conversational Interface — rather than navigating menus or writing queries, engineers interact with Sentinel in plain language. The platform interprets the intent and returns the relevant data, history, or analysis.
• Real-Time Data — Sentinel is designed to ingest live feeds from condition monitoring systems, keeping the asset model current and enabling time-sensitive queries against the latest state of the network.

Status

Sentinel is in active development. The graph data model and core architecture are established. The conversational query layer is being developed and tested against real asset datasets. Integration with condition monitoring feeds is planned for a later phase.

The goal is a platform that any engineer can open, ask a question, and trust the answer.