Architecture built to solve real problems.

The problem

In the UK, over 12,000 vulnerable adults lose housing every year due to missed correspondence deadlines. People with dementia, mental health conditions, or learning disabilities receive letters from the government, their landlord, or the NHS — written in legal language most people cannot understand. They contain deadlines that, if missed, can mean losing housing, losing benefits, or losing vital care.

Social workers and support workers try their best. But they're each managing 40+ clients at once. Letters pile up. Urgent cases get buried in routine post. By the time someone notices, the deadline has already passed. CivicGuardian was built because this keeps happening — and it doesn't have to.

What it does

CivicGuardian AI reads the letters. It identifies what is urgent, what the deadline is, what action is needed, and who needs to act — then surfaces a clear, prioritised alert to the right person before it is too late. It handles the triage. Humans make the decisions. A stack of 23 letters that would take a caseworker four hours to sort becomes a prioritised list in under eight minutes — with the most critical case flagged at the top.

Accepted document types

Tech stack

AgriSense — Australia's commercial farms are making critical decisions with incomplete data. That is what we are changing.

Australian broadacre farming runs on climate risk. Drought, frost, heat stress, and pest outbreaks are structural features of how the sector operates. A late October frost during wheat flowering can destroy an entire crop. A stripe rust outbreak undetected for 72 hours is an unrecoverable yield loss. The farms that manage this well have on-call agronomists and professional weather tools. The farms that don't are making the same high-stakes decisions with a regional forecast designed for an area 100 kilometres away — not accurate enough for field-level choices.

With A$101.4B in WA Wheatbelt sector output at stake, the intelligence gap matters. AgriSense delivers 7-day climate risk forecasts, pest and disease outbreak warnings with 72-hour advance notice, and planting window recommendations grounded in GRDC research — via a farm dashboard and REST API. At 5km-grid resolution.

Tech stack