Run a country on one operating picture — without a single ministry having to migrate.
Every ministry keeps its own system of record, its own ontology, its own data custody. Djeed reads each one, reconciles every entity, resolves the cross-agency record into one canonical knowledge graph — and from that foundation, every downstream capability becomes possible. Early warning. Cross-agency forecasts. Decision briefs traceable to source. Transparent automation. The audit trail an Auditor-General can read end to end.
The convergence
Many ontologies. One canonical knowledge graph.
Every ministry keeps the systems it already runs and the ontology its analysts already trust. Djeed reads what each one records, reconciles the entities across them — the same citizen, the same business, the same parcel, the same event — and resolves the cross-agency record into one canonical knowledge graph the whole government can navigate, query, and act on.
What it unlocks
Capabilities you couldn't build from one ministry alone.
Once the cross-agency record resolves into one canonical graph, every downstream capability runs from the same foundation. None of these are possible when each ministry can only see itself.
Early warning
Outbreaks, fraud, social-tension and supply signals seen in days, not in the next quarterly review. The signal lands when only one ministry sees it — Djeed correlates it before it becomes a crisis.
Cross-agency forecasts
Demand for clinics + schools + housing + roads in one projection. Capital planning that doesn't double-count and doesn't leave the gap that becomes next year's emergency.
Decision briefs
Cabinet- and Cabinet-Office-ready briefs assembled from records you can trace — not slides you have to trust. Every figure links back to its source ministry and source document.
Transparent automation
Routine cross-agency processes — permits, eligibility checks, benefit reconciliations — automated end-to-end, every step auditable. No black box, no shadow workflow.
Cross-agency search
A citizen, a parcel, a business entity, a piece of correspondence — one search across every ministry's record, returning everything that's relevant and nothing that isn't.
Audit traceability
Every figure in every report linked to its source record. The Auditor-General reads the same lineage trail as the analyst who wrote the brief.
Early warning · A real scenario
Eight days before the first hospital admission.
The same outbreak signal, surfacing across four ministries' data over eight days. On separate ledgers, each signal looks like noise. On the canonical graph, the pattern lands on day four — four days before the first patient is admitted.
School absences spike +35% in District 7
Ministry of Education
Signal · isolated
Fever-medication sales +180% same district
Health Ministry · pharmacy register
Signal · isolated
GP visits with ICD-10 J11 codes +50%
Health Ministry · primary care
Pattern caught
First hospital admission · respiratory failure
Health Ministry · hospital admission
Would-be · prevented
On day four, the cross-agency pattern lands on the canonical graph. The Ministry of Health gets a brief; the regional authority gets a heads-up; antiviral supply is checked against the projection. The hospital doesn't get blindsided on day eight — because someone could see the picture eight days earlier.
Connected to
Every ministry keeps the systems it already runs.
Djeed reads what's already there — without any ministry being asked to migrate, switch vendors, or hand over the primary system of record. Each ministry's existing ontology is preserved as a view; the canonical merge sits on top, not in place of.
The list opposite is illustrative — actual sources are scoped per government during a discovery engagement, against the data-protection and statistical- confidentiality posture of the country.
The five solutions, at national scale
The same flow. Sized for a country.
The five Djeed solutions are the same here as anywhere else — only the scale, the data sensitivity, and the deployment shape change.
Data & Ontology Foundation
Reads what every ministry already records. Reconciles their ontologies — `health.PatientRecord`, `civic.Citizen`, `fiscal.Taxpayer` — into one canonical model the whole government can act on.
Graph & Spatial Intelligence
The canonical knowledge graph, navigable across ministries. Every entity geo-tagged to administrative units — parcels, districts, regions — so cross-agency questions resolve to a place, not a spreadsheet.
Predictive & Generative AI
Forecasts grounded in actual cross-agency data — health × education × economic activity. Generative answers that trace to records, not models trained on the open web.
Operational Intelligence
Continuous monitoring across the whole national record. The signal in District 7 surfaces the day it starts — not in the next quarterly Cabinet paper.
Decision Intelligence
Cabinet briefs assembled from records you can trace. Decisions auditable from the figure on the page back to the source ministry's record. The Auditor-General reads the same lineage as the analyst.
For the technical reader — full per-solution detail at /solutions. The page above is the same five, framed for the government scale.
Sovereignty
Your country's data stays under your country's law.
For government, sovereignty isn't a footnote — it's the foundation. Djeed deploys air-gapped, on national infrastructure, on a sovereign cloud you control, or federated across each ministry's existing custody. The Swiss- hosted Djeed VM is one of many shapes, not the default.
Full air-gap available.
Djeed runs on your government's own data centre. No external network egress. The most sensitive deployments — defence-adjacent, classified records — run here.
Your hyperscaler region, your jurisdiction.
Hosted on a sovereign cloud you control — your country's law, your regulator's audit. The intelligence layer runs in a region you choose; data never leaves it.
Sensitive on-prem, intelligence in sovereign cloud.
Statistical, fiscal and citizen records stay at the ministry that owns them. The canonical graph runs in a federated layer above. Sensitive data never moves.
Each ministry keeps its own data.
The canonical graph runs as a federation across ministries — each retains custody of its records. Queries resolve across the federation; records stay where they were recorded.
Every deployment shape honours every applicable law of the country — data protection, statistical confidentiality, classified-data handling, archival law. We design the shape with your General Counsel and your data-protection authority before any data moves.
Trust
Every decision traceable. Every prediction explainable. Every exchange auditable.
Every record knows where it came from.
Source ministry, original document ID, capture timestamp, capture method — attached to every record. Not in a separate log; in the record itself.
Every prediction shows its work.
A forecast or a generative answer can be unfolded to the records it stood on. An auditor reading the brief can drill from the figure on the page to the row in the source ministry's database.
Every exchange is logged.
Every cross-agency data exchange logged immutably. The Auditor-General has the same view of the audit trail that the operator does — not a sanitised copy.
How a deployment starts
A pilot first, or the full programme. Either works.
Settle the shape, then expand.
Pick the ministry whose cross-agency questions are most urgent — typically Health, Statistics, or Finance — and run a focused engagement: one ministry connected, one canonical sub-graph live, one set of decision briefs in production. The deployment shape, the data-protection posture, and the governance model are settled before the rest of the government joins.
- One ministry connected, live in months
- Deployment + data-protection posture proven before expansion
- Each next ministry compounds on the foundation already laid
The whole government, phased.
For governments ready to commit to the operating picture from day one — a phased programme across every ministry, sequenced so the high-volume sources land first and the long-tail ministries join the canonical graph as their data is ready. Governance, audit and procurement framed at the centre, with ministerial leads inside each agency.
How a full programme is scoped →FAQ
Questions a Permanent Secretary asks first.
Does this mean we have to centralise our data?
+
Does this mean we have to centralise our data?
+No. The canonical knowledge graph is a layer above each ministry's existing system, not a replacement for it. Every ministry keeps its primary system of record, its existing ontology, and its data custody. Djeed reads and reconciles; it doesn't take ownership.
How does this comply with our statistical-confidentiality and data-protection law?
+
How does this comply with our statistical-confidentiality and data-protection law?
+Djeed is deployed in the shape your law requires. In jurisdictions where statistical records cannot leave the Statistics Office, the canonical graph runs federated — queries reach into the Statistics Office, but no record is exported. Same for fiscal records, judicial records, and other restricted categories. We design the deployment shape with your General Counsel and your data-protection authority before any data moves.
What happens to the ontology each ministry already uses?
+
What happens to the ontology each ministry already uses?
+Each ministry's ontology is preserved as a view on top of the canonical graph. Their tools, their reports, their primary system of record — unchanged. The canonical model is the cross-agency lens; the ministry's own ontology is the inside-the-ministry lens. Both stay valid.
Can the Auditor-General use the same system?
+
Can the Auditor-General use the same system?
+Yes. The audit role gets a read-only view of the same record + lineage that the operator has — not a sanitised copy. Every figure in every brief can be traced back to its source record by the auditor without going through the analyst.
How does the canonical knowledge graph stay accurate when ministries' data changes?
+
How does the canonical knowledge graph stay accurate when ministries' data changes?
+The canonical graph updates continuously from each ministry's source system — every record carries its own freshness timestamp, and the lineage to its source. Reconciliation is automatic where the ontology mapping is settled; edge cases route to a human reviewer with full source context attached.
What's the difference between a pilot and a full deployment?
+
What's the difference between a pilot and a full deployment?
+Pilot: one ministry (typically the one whose data is the messiest or the one whose cross-agency questions are the most urgent) live in months. Full programme: the whole government, planned in phases, sequenced so each ministry compounds on the foundation already laid. Most governments start with a pilot to settle the deployment shape and the data-protection posture before the wider programme.
Run your country on one operating picture.
Tell us about your government's structure and the cross-agency question you'd resolve first. We'll show you what the canonical knowledge graph would look like — and the sovereign deployment shape your law allows.