Phishing Tabletop Exercise: From Compromised Credentials to Crisis

One click, three decisions: a phishing tabletop exercise through lateral access, possible exposure, and a bank's 36-hour notification call.

A small signal branching through a fractured geometric network, representing phishing escalation.

Key Takeaways

  • A phishing tabletop exercise begins after the click and tests the institution's decisions once an attacker controls a working credential.

  • The exercise follows credential compromise, lateral access, and the determination of whether a notification incident occurred.

  • Under 12 CFR 53.3, the 36-hour clock begins when the bank determines that a notification incident has occurred.

  • The exercise should test that determination using the incomplete evidence available during the incident.

A phishing tabletop exercise tests what a bank decides after a credential is compromised. This walkthrough follows one treasury account through containment, lateral access, and the determination of whether a 36-hour regulatory clock begins. Three stages place progressively harder decisions in front of the same team.

What Is a Phishing Tabletop Exercise?

A phishing tabletop exercise is a discussion-based rehearsal that starts after an employee has already clicked, and walks a team through what the institution decides next. Its starting condition is what makes it a phishing variant: a working credential in an attacker's hands. It tests escalation, containment, and notification decisions. It does not test user awareness.

A phishing simulation measures whether employees click. A phishing tabletop exercise tests how the bank responds after a credential is compromised.

Click rate remains a useful control metric, but it does not test authority over a payment queue at 3:40 p.m. It also cannot show how long a decision-maker waits for confirmation the SOC cannot yet provide.

The APWG Phishing Activity Trends Report for the second quarter of 2026 counted 1,069,681 attacks, up 10.1% from the previous quarter. Financial institutions and payment brands together received roughly one-quarter of those attacks.

Reflex Security treats the gap between a compromised credential and an owned decision as the starting point for the exercise.

The broader exercise category and the reasons many programs change nothing are covered in why most tabletops fail. This article follows one escalation chain. A compromised credential becomes a session, and that session reaches a system of record. What it accessed informs the bank’s determination of whether the 36-hour clock begins.

How to Plan and Run a Phishing Tabletop Exercise

The room needs participants who can make decisions, not only describe them. Include the SOC analyst, incident response lead, legal counsel, communications, treasury operations, and an executive sponsor. The executive sponsor should have authority over customer-facing service availability.

The facilitator represents the primary federal regulator, a key commercial customer, and the cyber insurer. Keeping the same participants through all three stages exposes how ownership changes as the incident escalates.

Four artifacts support the exercise: the incident response plan, a recently verified escalation matrix, the notification rule, and an identity inventory. The inventory should identify which accounts can reach the CRM.

The FFIEC IT Examination Handbook's Business Continuity Management booklet places regulator, customer, and law-enforcement communications within incident response. Section VII.F also states that exercises should include communication processes with applicable stakeholders. The booklet is guidance rather than a binding requirement, but it supports the proposed cast.

The exercise should run for 90 to 120 minutes across three gated stages. The room advances only after stating the current decision and recording its timestamp. This structure prevents open-ended discussion from carrying the exercise forward without a decision.

How the injects themselves are written is a separate craft, covered in injects and escalation and in Cybersecurity Tabletop Exercises: Scenarios & Examples for 2026.

Phishing Tabletop Exercise Scenarios for Banking Teams

Five entry points can lead into the same three-stage chain:

  • Wire instruction change. A business email compromise message reaches the wire desk with an amended settlement account for a known commercial borrower. The credential that sent it belongs to the relationship manager, and it still works.

  • Vendor invoice fraud. A core banking vendor's accounts-receivable contact emails new remittance details from a lookalike domain. Accounts payable authenticates the request by replying to the same thread.

  • MFA fatigue. A treasury analyst approves the eleventh push notification at 11 p.m. to make it stop. The session is valid. The login is not.

  • Help-desk reset call. A caller with the CFO's date of birth and last four digits asks the service desk to move MFA to a new device. The desk follows its documented procedure and is still wrong.

  • QR-code lure on a shared device. A poster in a branch break room carries a payroll-update QR code. The tablet that scans it is shared, and no owner appears for it in the identity inventory.

A cyber attack tabletop exercise fails when each variant gets its own ending: five short discussions instead of one long consequence. Most published data breach tabletop exercise scenarios do exactly that. Tabletop exercise scenarios for banks have to carry a wire desk and a core banking platform, or the decisions cost nothing and the room knows it.

Sector programs can complement an institution’s internal exercises. FS-ISAC has run exercises since 2010 and describes Cyber-Attack Against Payments Systems (CAPS) as a tabletop team exercise for the payments industry. Versions cover banking, insurance, securities, and investments.

A sector exercise provides peer comparison. An internal exercise tests the institution’s reporting environment, wire desk, and regulator relationship.

Cyber Attack Scenario: An Employee Account Is Compromised

The scenario follows a regional commercial bank with roughly 2,000 employees. Its environment includes treasury operations, a wire desk, digital banking, a core banking platform, and a CRM containing customer PII.

Stage 1: Credential Compromise

A treasury operations analyst’s account authenticates from a residential IP in a state where the bank employs nobody. Forty minutes earlier, the same account authenticated from a branch.

The wire desk has six payment instructions queued for the 4:00 p.m. Fedwire cutoff. The analyst reviewed two of them. Who has authority to stop the queue, and does that person need confirmation from the SOC?

The room knows three things: one identity, two locations, and one successful MFA response. It does not know whether someone replayed a token or whether the session remains active. The analyst may also have approved a late-night push and forgotten by morning.

Each participant asks for something different. The SOC wants confirmation. Legal wants a timeline. Communications asks whether the incident is externally visible. The executive sponsor asks whether the Fedwire cutoff can move. Three requests seek information; one requires a decision.

Stage 1 decision

  • Owner: Incident response lead

  • Deadline: Before the Fedwire cutoff

  • Question: Is this an isolated credential compromise or a broader intrusion?

The data collector records the decision and its timestamp.

Stage 2: Lateral Access

The compromised session queries a shared reporting environment that combines core banking extracts with CRM exports. The query runs for nine minutes before the session becomes idle.

The environment has no data loss prevention agent, and its access logs roll every seven days. Who owns it, and when did anyone last confirm what it contains?

This transition is what the exercise needs to expose. A compromised credential became an active session, and that session reached a system of record. The incident may now involve data exposure, although the evidence does not yet prove exfiltration.

The SOC will probably detect the impossible travel and session anomaly. The harder test is whether the room escalates before confirmation arrives. The analyst may simply have logged in from a hotel.

Stage 2 decision

  • Owner: Executive sponsor

  • Deadline: Within 30 minutes

  • Question: Should the bank terminate the session and disable the account before confirming what the query returned?

Both choices carry consequences. Cutting access may disrupt online balance displays during business hours. Leaving it active increases the activity that investigators must reconstruct from seven-day logs. Teams unfamiliar with this decision often wait too long for forensic confirmation.

Stage 3: Possible Data Exposure and Notification Determination

Forensic review recovers a nine-minute query that returned 412,000 CRM rows. The records contain customer names, account numbers, and email addresses. The evidence neither confirms nor rules out egress.

Possible data exposure may trigger separate privacy, contractual, and state notification analyses. The federal banking rule asks a different question: did the incident materially disrupt, degrade, or threaten banking operations?

Under 12 CFR 53.3, the OCC must receive notification as soon as possible after the bank determines that a notification incident occurred. The deadline is no later than 36 hours after that determination. The Federal Reserve and FDIC adopted parallel requirements.

Section 53.2(b)(7) defines a notification incident through material disruption or degradation. The test considers banking operations, service to a material portion of customers, and material effects on a business line. A third prong covers threats to US financial stability.

The exercise must connect the CRM query to operational impact. The record count alone does not satisfy the federal banking rule. Uncertain egress also does not answer the operational question.

Stage 3 decision

  • Owner: General Counsel with the executive sponsor

  • Deadline: The 36-hour period begins after an affirmative determination

  • Question: Has the incident materially disrupted or degraded banking operations, or is it reasonably likely to do so?

Separate privacy, contractual, and state notification obligations belong in CISO Cyber Incident Response Plan: Prepare, Test & Respond.

Testing the Response: Detection, Escalation, Containment & Communication

Stage

The decision it forces

Owner

Clock

1 · Credential compromise

Is this one user, or a broader intrusion?

Incident response lead

Before the same-day Fedwire cutoff

2 · Lateral access

Does the bank terminate the session before confirming scope?

Executive sponsor

Within 30 minutes while the session remains active

3 · Notification determination

Has a notification incident occurred under 12 CFR 53.2(b)(7)?

General Counsel with the executive sponsor

36 hours after the determination under 12 CFR 53.3

The exercise should evaluate escalation rather than repeat the detection metrics already available in security dashboards.

Four failure modes can surface within 120 minutes:

  • The SOC raises the alert, but no defined action follows.

  • Escalation waits for confirmation and misses the window it was intended to protect.

  • Containment authority belongs to someone absent from the exercise.

  • A relationship manager receives a customer call without an approved response.

For each stage, the record should capture the decision, its owner, its timestamp, and the information available at that moment. Measuring tabletop results explains how to turn that evidence into corrective work.

This scenario requires conditions to change when the room stalls at a stage boundary. A fixed script can narrate that consequence, but it cannot generate a different path from the team’s decision. The distinction concerns exercise mechanics rather than the difficulty of phishing response.

Reflex makes the simulated adversary respond to participant decisions. If the team hesitates during Stage 2, the session can continue reading from the reporting environment.

Reflex uses One-Click Scenario Generation to produce five customized scenarios from a company domain in 10 to 15 minutes. Each Reflex exercise runs for 60 to 120 minutes with an AI facilitator named Victoria.

In Reflex, participants investigate simulated telemetry through the Investigation Console. The Facilitator Guide gives the session leader the questions to ask and the gaps to watch.

The Reflex After Action Report (AAR) typically exceeds 54 pages. It records each decision boundary with its timestamp and supporting evidence.

Reflex does not replace a knowledgeable facilitator who understands the institution. It provides an adaptive option when scenario supply, facilitation capacity, or scheduling limits the program.

Reflex explains the underlying mechanics in AI-driven injects and simulating a live crisis.

Book a demo to run the chain against the institution’s own environment with Reflex.

See what your team does under real pressure.

Put your readiness to the test. Sixty minutes, your real environment, no slides.

Book A demo

Frequently Asked Questions

How does a phishing tabletop exercise differ from a phishing simulation campaign?

A simulation sends test lures to employees and measures click and report rates. It is a controls program focused on user behavior. A tabletop assumes the click has already occurred and tests the institution’s subsequent decisions. The two formats measure different capabilities.

Should the exercise include a real simulated send?

It can, although a live send changes the exercise. It provides a genuine detection timestamp and real reporting behavior. It also adds scheduling, legal review, and the risk that discussion focuses on the message rather than the response. A first exercise can begin after the click.

Who from the business belongs in the room?

The exercise needs whoever can stop a payment queue and take a customer-facing service offline. At a bank, that usually means treasury operations and an executive with authority over online banking availability. Without them, the room may reach a decision it cannot implement.

How often should a bank repeat it?

A twice-yearly cadence allows the bank to repeat the decision chain with a different entry point. The second run can show whether decisions take less time and reach the correct owner without prompting.

What record should a phishing tabletop exercise leave behind?

No published standard defines one universal evidence package. A useful record identifies each decision, its owner, its timestamp, and the information available at that moment. An attendance sheet cannot provide that evidence.

References

APWG. (2026, August 28). Phishing activity trends report, 2nd quarter 2026. APWG. https://docs.apwg.org/reports/apwg_trends_report_q2_2026.pdf

Federal Financial Institutions Examination Council. (n.d.). Business continuity management: Incident response (§§ V.F.1 and VII.F). FFIEC IT Examination Handbook InfoBase. Retrieved September 10, 2026, from https://ithandbook.ffiec.gov/it-booklets/business-continuity-management/v-business-continuity-plan/vf-other-components/vf1-incident-response

FS-ISAC. (n.d.). Exercises overview: Cyber-Attack Against Payments Systems (CAPS). FS-ISAC. Retrieved September 10, 2026, from https://www.fsisac.com/hubfs/Resources/FS-ISAC_ExercisesOverview.pdf

Office of the Comptroller of the Currency. (2021, November 23). Computer-security incident notification requirements for banking organizations and their bank service providers, 86 FR 66442 (12 CFR pt. 53). eCFR. https://www.ecfr.gov/current/title-12/chapter-I/part-53

Cite this article

Cassio Goldschmidt.

Phishing Tabletop Exercise: From Compromised Credentials to Crisis.

Reflex Security.

https://reflexsecurity.com/guides/phishing-tabletop-exercise-from-compromised-credentials-to-crisis

Incident closed // after-action ready

Incident closed

See what your team does

under real pressure.

True cyber crisis readiness isn't built in a slide deck—it's forged in the trenches. Every team believes it’s ready. The simulation is where you find out. Sixty minutes, your real environment, no slides.

Book A demo

ReflexSecurity

AI-powered incident-response simulation platform.

AI-powered incident-response simulation platform.

© 2026 Reflex Security. All rights reserved.