The weak signal caught before the crisis
Communities often know a crisis is coming before the services meant to respond to it. They see the river rising, they feel the tension shift, they open the same section again and again. The problem is not a lack of signals, it is that they do not reach the right people in time. The SOLIS bot's early-warning system catches these weak signals and pushes them, in real time, to your teams.
How it works
The bot quietly watches for three kinds of signal:
- a spike in consultations of one section, for example dozens of people opening "evacuation routes" at three in the morning,
- converging reports from several community relays about the same event,
- an abnormal volume of incoming messages in one area.
When a threshold is crossed, your teams receive a notification, targeted on the topics you have chosen to monitor: water, food security, protection, MEAL. A real example, in Cameroon's Logone valley:
[ALERT] 3 reports of rising water received this morning in the Logone-Birni area. Villages: Doumrou, Kousséri-Rural, Fotokol-Nord. See the dashboard for details.
One hour after this notification, once validated, 847 households in the area received the flood alert.
The non-negotiable principle: the bot never decides alone
Every alert goes through human validation, following an escalation protocol written in advance: who validates, how fast, who triggers the broadcast. In the reporting forms, a question "is anyone in immediate danger?" triggers, on a positive answer, a direct notification to the programme coordinator for human review. Technology detects, humanitarians decide. Always.
The network that makes it possible
The strength of the system lies in its community focal points. Trusted relays, identified in each village, are given a specific role in the bot that opens dedicated reporting forms to them, invisible to ordinary beneficiaries. They become, in our teams' words, "your most valuable field sensors": eyes where no team can go, around the clock. A network of thirty relays sending eight reports out of ten beats a network of five perfect but isolated ones.
What it changes
Early warning moves the line between enduring and anticipating. Where field information took days to travel, when it travelled at all, it now reaches teams within minutes, before an officer is even aware. In a disaster-preparedness context, those few hours gained can change the outcome of a rainy season.
For a system that holds
- Document the escalation protocol in writing, before the risk season. Who validates the alert, how fast, who broadcasts if the focal point is unreachable.
- Calibrate report frequency to the context: weekly in calm periods, every two days in the rainy season. The bot lets you adjust the reminders in a few clicks.
- Train a backup relay in each village. If the main focal point loses their phone or is away, the network must not collapse.
- Fine-tune the thresholds after two weeks. By then you will have a sense of "normal" consultation volume, and you can set the alert threshold precisely.