Early Warning in the Sahel

The Challenge

The Sahel has become the most coup-prone region on earth. Mali in 2020 and again in 2021, Burkina Faso twice in 2022, Niger in July 2023, with jihadist insurgency spreading, the Wagner and Africa Corps footprint expanding, the Western security presence drawing down, and the regional order fracturing as Mali, Burkina Faso, and Niger formed the Alliance of Sahel States and moved to quit ECOWAS. For an operator with fixed, immovable assets and thousands of staff and contractors on the ground, this was an existential operating environment. The client’s posture was reactive. It learned about deterioration roughly when the international press did, by which point borders were closing, airspace was restricted, and the window to protect people and property had already shut. The asset was good. The decision loop was too slow.

The Approach

We built the engagement explicitly as an observe-orient-decide-act loop, because in a fast-moving security environment the side with the faster loop wins, and here the adversary was the crisis itself. Observe meant replacing headline-watching with a structured, country-specific indicator set: troop and unit movements, signs of elite or military defection, fuel and cash shortages, widening spreads in the parallel currency market, shifts in local-language social-media sentiment, and road and checkpoint closures, fed by a local stringer network as well as open sources. The leading signals of instability are rarely in English and rarely in the news. Pre-emptive drawdown executed five days before borders and airspace closed More than 400 staff and contractors protected; $250m in assets secured Insurance premiums reduced by 18% on the strength of the new posture

Orient stood up a small analytical cell to turn raw signals into assessments, weighting them against the historical patterns that precede coups and escalations in the region, and mapping them to defined scenarios with the specific indicators that distinguish one from another. Decide was where most crisis plans fail, and where we spent the most effort. We agreed trigger thresholds in advance, each tied to a specific action: a travel freeze at one level, a phased drawdown at the next, full evacuation and asset safe-mode at the most severe. Decision rights and the call-tree were assigned before any crisis, so that when a threshold was crossed nobody was relitigating who had authority while the situation moved. Act made the plans real through rehearsal and pre-positioning: evacuation routes and charter options identified, fuel and communications staged, playbooks drilled rather than filed. Three things made the system more than a dashboard. First, it was built to beat the CNN test, the failure mode where a company learns about a coup when its own staff see it on the news; every indicator earned its place by moving before the headline, not with it. Second, we tied the posture to the contractual and insurance architecture, so that a drawdown triggered the right force-majeure notifications and kept war-risk cover valid rather than voided. Third, and most consequential legally, the system gave the company a defensible duty-of-care record: when an operator moves staff out of harm’s way ahead of an event, on documented criteria, it is discharging an obligation that courts and regulators increasingly expect and that peers who improvise cannot demonstrate. The analytical cell was kept small and senior rather than a large watch floor, because the bottleneck in these situations is judgment, not screen coverage.

The Outcome

Ahead of one episode, the system flagged the deterioration early and the client executed a pre-emptive drawdown five days before borders and airspace closed, protecting more than 400 people and securing $250m in assets while peers were caught flat. The improved posture also cut insurance premiums by 18%. The client kept operating, on its own timetable, in a region where most of the field was reacting.