Exercise NORTHERN LANTERN: What Joint Intelligence and Policing Looks Like at Speed

Training scenario, illustrative only. Exercise NORTHERN LANTERN is a hypothetical scenario written for training and capability demonstration. It does not describe a real operation, customer or dataset.

Ask anyone who has worked a joint intelligence and policing problem where the time actually goes, and you will rarely hear “analysis”. You will hear collation. Chasing a spreadsheet. Re-keying a manifest. Waiting for a partner to confirm whether two differently spelled names are the same person. By the time the picture is assembled, the decision window has often closed.

We designed Exercise NORTHERN LANTERN as a hypothetical five-day scenario to test a different rhythm, one where getting information in and producing intelligence from it are treated as two distinct, well-engineered jobs.

The set-up

A fictional regional police force, a national intelligence customer, a counter-fraud regulator and two authorised partners work a developing threat picture around a logistics hub: access-control anomalies, a cluster of suspicious company registrations, haulage manifests that do not reconcile, and rising open-source chatter. Every team holds part of the answer. No team can lawfully hand everything to everyone.

NERVE gets information in

Each team contributes only what it is authorised to contribute, in whatever format it already holds, documents, scans, exports, structured feeds, imagery. NERVE indexes it for governed search while leaving the originals untouched, and carries source, contributing organisation, collection time and handling marking with every item. Integration is service-mediated, so contributing systems are never merged or exposed to one another.

Candidate entity matches across spellings, registrations and vehicle identifiers are proposed to analysts for confirmation rather than quietly merged. Gaps stay visible, which matters: an analyst needs to know what is missing as clearly as what is present.

CORTEX turns it into intelligence

Assessments are built against a structured template with confidence and reliability recorded explicitly, and every claim stays linked to the material behind it. When day four brings a corroborating partner sighting, the assessment is revised and reissued rather than rewritten, with the change history preserved. From one evidence base the team issues three outputs: a decision brief, a fuller analytical report and a tasking summary.

Sharing without over-sharing

This is where joint working normally breaks. The scenario tests deliberate release instead: a named releasing officer, a recorded justification, a defined recipient group, and product kept separable from material so a partner can receive an assessment, the supporting material, or a redacted subset of either. Caveats travel with the release rather than sitting in a covering email. A partner can see that a judgement rests on corroborated material without ever being handed the protected source.

What the exercise demonstrates

• Effort moves from collation to judgement.

• Assessments are produced in hours and revised as material lands.

• Partners act on product without holding material they are not cleared for.

• Post-exercise review is evidential, because ingest, query, production and release were all logged as they happened.

The full scenario walkthrough, including the day-by-day timeline, is written up as a case study.

If you would like us to design and facilitate a tailored scenario against your own operating model, get in touch.

Explore the platforms: nerve.aic.io and cortex.aic.io.