
JDE Sentinel™
ERP monitoring and anomaly detection Agent
Part of JDE AgentOS™ · Governed. Reliable. Built for JD Edwards.
- Type
- AI Agent
- Platform
- JD Edwards
- Category
- Support & operations
- Governance
- Human approval before any action
The Problem It Solves
Who works with it
- Managed services team
- JD Edwards administrators
- IT operations
By the time someone notices that a nightly batch schedule did not finish, an integration queue has been backing up since midnight, or a server is running out of headroom, users are already affected and the month-end is at risk. Dashboards show the numbers; nobody is watching them at 3 a.m.
JDE Sentinel™ watches continuously. It learns what normal looks like for your estate (job durations, queue depths, error rates, login patterns) and raises the exceptions that matter, with the likely cause and the recommended action, before they become incidents.
What JDE Sentinel™ does
Monitors batch schedules, subsystem jobs and orchestrations for failures, delays and unusual run times
Watches integration queues, AIS calls and interface tables for backlog and error spikes
Tracks performance and capacity signals across the enterprise and database servers
Flags unusual access patterns and security-relevant configuration changes
Groups related symptoms into one alert with a probable cause instead of a flood of notifications
How JDE Sentinel™ works
The same governed loop as every Agent on JDE AgentOS™ (monitor, analyse, recommend, execute), applied to this Agent's job. Nothing reaches JD Edwards without a person approving it.
Monitor
Collects job, queue, performance and security signals from JD Edwards around the clock.
Analyse
Compares them against learned baselines and known failure patterns to spot anomalies.
Recommend
Raises a prioritised alert with the probable cause and the suggested action.
Execute
Restarts or re-queues only what has been pre-approved; anything else goes to the on-call team.
What Data It Touches
- Reads: job and subsystem logs, orchestration and integration queues, server metrics, security logs
- Writes: alerts, incident notes and pre-approved restart or re-queue actions
- Never touches: business transactions or master data
Security: Where the Data Stays and Who Can Access It
A governance layer routes every action through human approval workflows, and the action layer only executes JD Edwards transactions that have been approved. The connector layer is AIS, Orchestrator, REST APIs and the database.
Where it sits in the JDE Agent Fabric
JDE Sentinel™ lives in the Agent layer. Everything above it governs what it may do; everything below it is how it reads JD Edwards.

01Action layer
Execution of approved JDE transactions

02Governance layer
Security, audit, human approval workflows

03Agent layer
JDE Sentinel™ and the rest of the portfolio

04Reasoning layer
LLMs, RAG, rules, business policies

05Enterprise knowledge layer
Documentation, SOPs, configurations, historical tickets

06JDE connector layer
AIS, Orchestrator, REST APIs, database
Related
- Service
JD Edwards services
Implementation, upgrades and support on JD Edwards.
Explore →
AI AgentJD EdwardsSupport & operations
JDE Doctor™
AI-powered support and issue resolution assistant
See how it works →
AI AgentJD EdwardsUpgrades & quality
JDE UpgradeIQ™
Upgrade, customization and impact analysis assistant
See how it works →
See JDE Sentinel™ on your own data
A short demo against your environment, with your team in the room.

