Monitor
Social-media monitoring
A purpose-built, state-of-the-art layer that watches public social platforms and channels and scales to any volume — surfacing the activity that looks like a scam.
Public safety · Anti-scam
Anti-scam detection
Scams don't operate alone — they run as rings, scattered across public social platforms. ASTRA's anti-scam system monitors those signals at scale and relates them into a graph, so the whole network is exposed — not just the single scam.
For a Hong Kong public-safety agency.
The challenge
The challenge
Scams move fast, and the signal that exposes them is scattered across public social platforms and channels. Worse, scams rarely operate alone: they run as rings of connected accounts, numbers and links. Spotting one scam post is hard enough; piecing together the network behind it — by hand, at scale — is effectively impossible. The connections that reveal the ring stay invisible, and the people who could act are left chasing isolated reports.
How we reframe it
How we reframe it
ASTRA built an anti-scam intelligence system on two purpose-built foundations. A state-of-the-art social-media monitoring layer watches public social platforms and channels and scales to any volume, surfacing the activity that looks like a scam. A graph engine then relates those signals — linking the accounts, numbers, links and the connections between them — to expose the scam ring behind the activity, not just the isolated incident. The result reaches a Hong Kong public-safety agency as specific, actionable intelligence. The methods draw on research-grade scam-prevention work from the ASTRA / AIFT lineage. It is a developed, working system, proven on real data — and the same foundation, monitoring plus a graph engine, is built to power other intelligence systems for other purposes, with anti-scam as the first. A new deployment begins with a short internal integration trial in the client's own environment.
The expected outcome
The expected outcome
The aim is faster, better-targeted action against scams — by exposing whole networks rather than chasing isolated reports, and handing a public-safety agency the connected picture it needs to act. Because it sits on a general intelligence foundation, the same monitoring-and-graph approach extends well beyond anti-scam.
- Whole scam rings surfaced — not isolated incidents
- Public social signals related into one connected picture
- A general intelligence foundation that extends beyond anti-scam
The engine
The engine behind it.
Two purpose-built foundations do the heavy lifting — and the same pairing is built to power intelligence systems well beyond anti-scam.
Relate
A graph engine
Relates the signals — linking accounts, numbers, links and the connections between them into one graph — to expose the scam ring behind the activity, not just the single scam.
How it works
Signal in, action out.
Signals
Public social signals
Public social platforms and channels, monitored at scale.
Monitor
Social-media monitoring
Purpose-built infrastructure surfaces the activity that looks like a scam.
Relate
Graph engine relates the data
Linking accounts, numbers and connections to expose the ring behind them.
Act
Flagged to the agency
Specific, actionable intelligence for a public-safety agency.
Built for a Hong Kong public-safety agency, on real data — the partner is described in general terms. The scam-prevention methods come from the ASTRA / AIFT research lineage, work led by Prof. Xiaofan Liu.
Across the ASTRA team and the AIFT research lineage. Client names available on request.