CrimsonRadar: Device security case study

7 views

from fleet posture to incidents and patching, all working on one shared set of sample devices

Scope fact

3 presets

Standard, High-Risk and Travel lockdown, each a set of hardening controls applied to a device group

Design decision

4 factors

behind every area score on the globe, each shown with its weight, so a rating always comes with its reasons

Design decision

The problem

Phones are an easy way into a company: a phishing text, a Wi-Fi network that copies the office name, an exploit that needs no tap at all. IT teams see these in separate tools and have to piece them together while the threat is still live.

CrimsonRadar puts device posture, detections, networks, hardening and response on one console, and extends the same idea to people on the move: a globe that shows how safe an area feels right now, and why.

Key decisions

Constraint

Hardening a phone means dozens of settings, and IT teams rarely have time to tune them person by person.

Options

Expose every control, or start from presets.

Decision

Lockdown presets: Standard for everyday use, High-Risk for executives, legal and press staff, and Travel for hostile networks. IT can fine-tune each one and apply it to a device group.

Trade-off

Presets are opinions. Every control inside them stays visible and adjustable.

Constraint

A detection only helps if someone acts on it before the damage is done.

Options

Alert and wait for a person, or take the safe first step automatically.

Decision

Automatic first steps, such as isolating the device, blocking the connection or forcing the VPN, shown on the detection, with an incident opened for a person to work.

Trade-off

An automatic step can interrupt someone’s day. Each one is recorded with what triggered it.

Constraint

People only act on a risk score they trust.

Options

Show one number, or show the number and its reasons.

Decision

Every area on the globe opens its breakdown: reported incidents, weather, crowds and lighting, each with its weight, next to a fastest-versus-safest route.

Trade-off

More to read. It is also what makes the number worth following.

The work

CrimsonRadar dashboard with open threats, devices at risk, patch coverage and a seven-day threat trend — **Dashboard**Open threats, devices at risk, patch coverage and average risk across the fleet, with the week’s threat trend and today’s threat categories. Demo data

CrimsonRadar device posture list with owners, risk scores, patch levels and compliance — **Device posture**Every phone, tablet and laptop with its owner, risk score, patch level and compliance. Demo data

CrimsonRadar threat detection with a live threat feed, threat types and a behavioural anomaly chart — **Threat detection**A live feed of zero-click attempts, phishing links and rogue networks, with the action taken automatically and the confidence of each detection. Demo data

CrimsonRadar lockdown modes: Standard, High-Risk and Travel presets with hardening controls — **Lockdown modes**Pick a preset, fine-tune its hardening controls, then apply it to a device group. Demo data

CrimsonRadar incidents list with severity, owner and status — **Incidents**Each with its timeline, owner and notes. Change the status, assign it or add a note in place. Demo data

CrimsonRadar Live Risk Globe zoomed to an area scored 6.5 out of 10, with the factors behind the score — **Live Risk Globe**How safe an area feels right now, and the four weighted factors behind the score. Made-up places.

CrimsonRadar safe route planner comparing a fastest and a safest route — **Safe route planner**The fastest route against the safest one, with the risks along each. Made-up places.

How it was built

01DiscoverCalls and walk-throughs with the people who would use it
02CompileTranscripts, files and constraints in one requirements brief
03SpecifyA detailed, sourced spec from Perplexity, reviewed and locked
04BuildDesigned and built with the latest AI models, in the production stack
05TestUser testing on every preview, with real users on real tasks
06IterateFindings turned into the next round of changes, worst problems first
07ShipTests, security review and release, then back to testing and iterating↻ Back to Test
↻ Test and iterate again

How I work, step by step →

Where AI was used

Built with coding agents as a clickable prototype. Every company, person, device, detection and vulnerability ID is made-up sample data, and nothing touches a real device.

What’s next

Show it to IT and security teams to find which views they would use first, then decide whether to connect it to a real device-management system.