Corporate Social Engineering
Targeted phishing campaigns, pretexting, vishing, and physical access testing tailored to your organisation. We expose the human vulnerabilities that no firewall can patch and help you build a security-aware culture.
Bespoke phishing simulations, pretexting calls, physical tailgating assessments, and USB drop tests. Full reporting with staff awareness metrics and recommended training paths.
What we cover
Phishing
Targeted email campaigns built from real reconnaissance, not a generic template with your logo pasted in.
Vishing
Telephone pretexting against helpdesks and support staff, including password and MFA reset flows.
Physical access
Tailgating, reception pretexts, and whether an unaccompanied visitor can reach a live network port.
Device drops
USB and QR payloads left where staff will find them, measuring what gets picked up and connected.
Reporting culture
How quickly suspicious contact is escalated, and whether staff who report it get a useful response.
Technical controls
The email authentication, filtering and endpoint protections that should have stopped us before anyone was tested.
Your firewall cannot be talked into making an exception. Your staff can, and that is not a failing on their part. People are employed to be helpful, to move quickly, and to solve the problem in front of them. A well-built pretext turns each of those virtues into an attack path.
Social engineering testing is only worth doing if it produces a change in process rather than a list of names. We design our engagements accordingly.
How we approach a social engineering engagement
We start with open-source reconnaissance: what your organisation discloses publicly about its people, its suppliers, its tooling and its internal vocabulary. That research is what separates a credible pretext from an obvious one, and it is also a finding in its own right. Clients are routinely surprised by how much of their internal structure is externally visible.
Campaigns are then built specifically for your organisation. We agree scope, timing and escalation contacts in writing beforehand, including which teams are in scope and what we will do if a test produces genuine distress or a real security incident.
How we report
We report at the level of process and control, not individuals. Click rates are given as aggregate figures, because the actionable finding is almost never "this person clicked a link". It is that a convincing message reached an inbox at all, that there was no obvious way to report it, or that a helpdesk reset a credential on the strength of a confident phone call.
Where our pretexts succeeded, we explain which control should have stopped them and what to change. Where staff caught us, we say so: that is the result worth reinforcing.
The same problem, pointed at a machine
Prompt injection is social engineering aimed at software: a convincing instruction from an untrusted source, arriving through a channel the system was built to treat as trustworthy. The mitigations rhyme too. You do not fix either by expecting the target to spot a malicious instruction. You fix both by limiting what acting on one can achieve.
If you have given a language model access to internal systems, our AI and LLM application security work tests that limit.
What you get
01
Campaign metrics covering click, submission and reporting rates
02
A written report that identifies process failures rather than naming individuals
03
Recommended training paths targeted at the weaknesses actually found
04
A re-test after training to measure whether anything changed
Other services
Every engagement starts with a conversation about scope. Tell us what you have built and what worries you about it, and we will tell you what a useful assessment looks like.