Social Engineering Assessment (OSINT-Led)

An OSINT social engineering assessment tests whether the details anyone can find about your organization would let an attacker talk their way past your processes. We build realistic phishing, phone and process scenarios from your actual public exposure, run them only with written authorization and stop rules, and report on what failed in the process, not who clicked.

  • Scenarios from real public exposure
  • Written authorization and stop rules
  • Processes tested, not people blamed
  • Results by team, not by name
Short answer

An OSINT social engineering assessment uses open-source reconnaissance to design believable attacks on your people and processes, then tests them under written authorization: phishing emails, calls to the help desk, supplier payment changes. OSINT-S agrees targets, exclusions and stop rules in advance, never captures real passwords and reports results by team and process, so the outcome is fixed controls rather than blamed staff.

What an OSINT Social Engineering Assessment Tests

Whether a request built from public information gets through a business process. The person who answers is part of the process, not its weakest link.

Generic phishing templates measure whether staff recognize generic phishing. Attackers do something else: they read your website, job ads, press releases and staff profiles, then write a message about your real supplier, your real software migration or your real office move. An OSINT-led assessment copies that method, so the result tells you how your controls hold up against the attacks you are likely to receive.

The question in every scenario is the same: did the process stop it? A finance clerk who pays a convincing fake invoice has exposed a missing call-back check, and a help desk agent who resets a password for a confident caller has exposed a weak identity check. Those are the findings that get fixed.

OSINT-Built Scenarios and the Controls They Test

Each scenario starts from something we found in public sources and ends with a named control that either worked or did not.

ScenarioBuilt fromControl tested
Help desk callStaff names and roles, help desk number, email formatIdentity checks before password or MFA resets
Supplier bank changeSuppliers named in case studies, tenders or pressCall-back to a known number before payment details change
Targeted phishing emailCurrent projects, software in job ads, recent announcementsEmail filtering, reporting button, sign-in protections
Executive messageExecutives' public travel and speaking scheduleVerification of urgent requests made outside normal channels
Visitor entryOffice photos, badge designs, delivery routinesReception and access procedures, only with the site owner's permission

The recon behind these scenarios is the same work described under red team OSINT reconnaissance; the assessment adds the controlled test and the debrief.

Authorization, Scope and Stop Rules

Every scenario, target group and exclusion is written down and signed before the first message is sent.

  • Sign-off. A senior sponsor with authority over the people and processes in scope, with HR and legal aware of the test. For physical entry, the site owner and, where relevant, the landlord also agree.
  • Targets and exclusions. Named teams or roles in scope; people on leave, under investigation, recently bereaved or otherwise vulnerable are excluded, and so are personal phones and personal accounts.
  • Off-limits lures. No messages about pay, bonuses, layoffs, health, family emergencies or disciplinary action. No impersonation of real third parties, such as a named supplier, bank or government body, without their written consent.
  • No real credentials. Test pages record that a submission happened, not what was typed; nothing entered is stored.
  • Stop rules. A named contact who can halt the test at any time, and an immediate stop if a scenario causes distress, reaches someone outside scope or triggers a real incident response.

Employee Welfare and a No-Blame Approach

Tests that punish people make reporting worse. We design the assessment so staff come out of it more willing to report, not less.

The UK's NCSC, the government's cyber security authority, warns that blaming users for clicking on links does not work, that users who fear reprisals will not report mistakes promptly, if at all, and that phishing simulations can erode trust between employees and security and even create legal risk. It also accepts that some phishing attacks will always get through (NCSC). Our design follows from that:

  • Anyone who falls for a scenario sees a short, friendly explanation at once, not a warning.
  • Results are reported by team, role and process; individual names are not shared unless your written scope says otherwise for a specific process owner.
  • People who report the test are recognized, and report rate and time to report are measured alongside click rate.
  • Findings go into training and process changes. NIST's guidance on learning programs focuses on behavior change and a strong security culture rather than one-off tests (NIST SP 800-50 Rev. 1).

How the Assessment Runs

Recon, scenario design, sign-off, a controlled test window, debrief and a re-test after fixes.

  1. ReconOpen-source collection on your organization's public exposure, limited to what the scenarios need.
  2. Design scenariosThree to six scenarios, each tied to a control, with the public evidence that makes it believable.
  3. Sign-offSponsor, HR and legal approve scenarios, targets, exclusions, dates and stop rules in writing.
  4. Test windowScenarios run in an agreed window, with a live contact on your side and daily status notes.
  5. DebriefA no-blame session with process owners, then a short message to all staff explaining what happened and why.
  6. Re-testAfter fixes, the weakest scenarios are repeated to show whether the control now holds.

Deliverables, Metrics and Timelines

A report on which controls held, aggregated metrics and a fix list, from 10 business days for a focused assessment.

You receive the scenarios with the public evidence behind each, results by control and team, report rates and times, and a ranked list of process fixes, plus the anonymized lures for your awareness program. A focused assessment of two or three scenarios runs from 10 business days; a broader program across sites or brands takes up to about a month. Urgent delivery costs 50% more, the price is fixed after written scoping, a senior analyst reviews the report and the fee goes down if we miss the agreed date. Staff data is minimized and deleted at the end as agreed under the GDPR and similar laws.

If you want the technical side tested too, pair this with OSINT penetration testing recon. Real lures already aimed at your staff can be studied through phishing intelligence, and executives' personal exposure through a personal digital footprint check. Social engineering testing is one of the OSINT services for security and awareness teams we run.

Test the Process, Not the People

Send the teams and processes you want tested, any scenarios you already have in mind and who will sign off. We propose scenarios, rules and a fixed price.

OSINT Social Engineering FAQ

Our annual phishing simulation uses generic templates and almost nobody clicks anymore — would an OSINT social engineering assessment show us something different?

Very likely. Generic templates mainly test whether staff recognize generic phishing. An OSINT social engineering assessment builds scenarios from what attackers can actually find about you, such as named suppliers, current projects and staff roles, and tests the process behind each one, like payment changes or help desk resets. The result shows which controls hold against attacks you are likely to receive.

I'm the CISO and our HR director worries that tricking staff will damage trust — how do you run social engineering tests without people feeling set up?

By agreeing exclusions and off-limits lures with HR before anything runs, giving anyone who falls for a scenario an immediate, friendly explanation, and reporting by team and process rather than by name. People who report the test are recognized. A short all-staff message after the test explains what happened and why. HR signs off on the plan, so there are no surprises.

Can you call our help desk pretending to be one of our executives to see whether they will reset her MFA?

Yes, if it is in the signed scope and the executive has agreed to her name being used. The call follows an approved script and ends the moment the agent asks for verification we cannot provide, or performs the reset, which we then record and help reverse. We never use the reset to log in. Calls to personal numbers are out of scope.

We want to send a test email pretending to be our biggest supplier asking for a bank detail change — is that allowed?

Only with the supplier's written consent, because the test impersonates a real company. Without it, we build a scenario around a fictitious but plausible supplier and test the same control: whether finance calls back on a known number before changing payment details. Either way, no real payment is made and the test stops at the point where details would be changed.

Will your report name the employees who clicked or gave information away, and can we use it in their performance reviews?

By default, no. Results are aggregated by team, role and process, because the aim is to fix controls. Using test results for discipline discourages reporting, which the UK NCSC also warns against. If your scope names a specific process owner, that is agreed in writing in advance. We recommend keeping assessment results out of performance reviews.