- 01
Market
Saudi Arabia
- 02
Written for
CISO
- 03
Decision
Improve detection coverage through joint attacker-defender exercises
- 04
Languages
EN · AR
The working loop
One technique, run and re-run until the alert fires
The unit of work is not a campaign but a loop, repeated for each technique on the list. Its value is the short distance between an attack step and the fix for missing it.
A loop of four steps — choose the technique, execute it live, read what the telemetry saw, tune and re-run — with a branch recording gaps that tuning cannot close into an engineering backlog.
- 01
Choose the technique
Drawn from the behaviours relevant to your sector and estate — informed by intelligence where you have it — not from an alphabetical walk through a framework.
- 02
Execute it live
Our operator runs the real technique on agreed systems, announced and scheduled — the point is producing genuine telemetry, not surprising anyone.
- 03
Read what the telemetry saw
Defenders trace the execution through their own logs and sensors: was the event recorded at all, did a rule match it, did an alert reach a human.
- 04
Tune and run it again
The rule is written or sharpened in the session, and the technique is re-run immediately — the pass is not a promise to fix but the alert actually firing.
Record the gap tuning cannot close
Some misses are not rule problems — the log source does not exist, or the architecture hides the event. Those go to an engineering backlog with the evidence attached, which is a finding as valuable as any detection fixed.
Against the red team
Six weeks of silence, or an answer the same afternoon
A red team and a purple team use the same tradecraft to answer different questions. The red team is adversarial and quiet: it asks whether a determined attacker could reach a specific objective, so it hides from your defenders and reports at the end — and it is measured by objectives reached. The purple team is collaborative and loud: it asks how much of the attacker's playbook your detection actually sees, so it announces every move and is measured by detections improved. One tests the outcome; the other manufactures the capability.
The choice between them is mostly a question of what you would do with the result. An organisation whose detection coverage is thin gets little from a red team beyond confirmation — the operators reach the objective, the report says what everyone suspected, and the fixing still lies ahead. The purple team spends the same offensive skill directly on the fixing: every hour of attacker time becomes tuned rules, validated log sources and analysts who have now watched real techniques traverse their own screens. When coverage has been built and the question becomes whether it holds against an adversary who is genuinely trying, that is the red team's moment.
A typical sprint
Two weeks that turn a coverage guess into a coverage map
Illustrative engagement, not a client case
A security team has invested in a SIEM and an endpoint platform and believes its coverage is reasonable, but the belief rests on vendor claims and default rules — nobody has watched a real technique cross the estate and checked what fired.
- 01A technique list is agreed in advance — a few dozen behaviours spanning credential abuse, lateral movement and exfiltration, weighted toward what targets the sector — with systems in bounds and a stop authority named.
- 02Each day runs the loop: techniques executed, telemetry read together, rules tuned, re-runs until the alert fires — with defenders driving their own consoles, because the skill has to land in their hands, not ours.
- 03Misses that tuning cannot fix — an absent log source, an unmonitored segment — are recorded with evidence and costed into an engineering backlog rather than hidden in an appendix.
The team ends the sprint with a technique-by-technique map showing what was caught before, what is caught now, and what needs engineering — a before-and-after the CISO can put in front of leadership, and a loop the team can keep running on its own.
Our role
Which chair we take in the exercise and what your SOC keeps
- 01
The attacker's chair
We bring the offensive operators and the technique tradecraft, executed safely on agreed systems — the half of the exercise most internal teams cannot staff themselves.
- 02
Detection engineering beside yours
Our detection engineers sit with your analysts during tuning — suggesting the rule logic, explaining what the technique leaves in the logs — while your team writes and owns every rule that results.
- 03
The map and the backlog
We maintain the coverage map through the sprint and hand it over with the engineering backlog priced and ordered, so the exercise ends in a plan rather than a memory.
The detections, the map and the practised loop stay with your SOC; when the next sprint runs, it starts from where this one finished rather than from the beginning.
Common questions
What buyers ask
Do we need a mature SOC before purple teaming is worth the money?+
You need less maturity than for a red team, and more than none. The exercise requires two things to exist: telemetry that records something — a SIEM or endpoint platform collecting from the systems in scope — and at least a small team who will own the rules being tuned. Given those, low maturity is an argument for purple teaming, not against it, because every session directly builds the capability whose absence is the problem. What the exercise cannot do is substitute for the platforms themselves; if nothing collects the logs, the first finding will be that, and the engagement becomes an engineering plan rather than a tuning sprint.
What does progress look like across repeated purple team sprints?+
Three curves, all visible in the coverage map. First, the catch rate on first execution rises — techniques fire an alert before any tuning, because earlier tuning generalised. Second, the time from execution to tuned detection falls, because your analysts have done the loop before and know their own telemetry. Third, and most telling, the findings migrate: early sprints surface rule problems, later sprints surface architecture problems — missing log sources, unmonitored segments — which means tuning has stopped being the constraint. When a sprint produces mostly engineering items and few rule fixes, the programme has done its job, and the question becomes when to bring in a red team to test the whole under adversarial conditions.
Can purple teaming run with our MDR provider instead of an internal SOC?+
Yes, and it is one of the most useful configurations, because it converts a vendor relationship into a measured one. The provider's analysts join the sessions, the techniques run against the estate they monitor, and everyone sees — on the provider's own consoles — what was caught, how fast, and what needed tuning on their side. That replaces the standing question of what a detection service actually detects with a technique-level answer, and the coverage map becomes the factual basis for the next service review. Providers confident in their service generally welcome it; reluctance to participate is itself information.
In this cluster
Next step
Plan a purple team sprint
A scoped first sprint: the technique set chosen from threats relevant to you, the systems in bounds, and the coverage map both teams will fill in together.

