Blog
tabletop exercise scenariosincident response tabletop exercisetabletop exercise template

Tabletop Exercise Scenarios for Incident Response (Template)

Ready-to-copy tabletop exercise scenarios for incident response: a full template, a worked example, and AI facilitator prompts for audits and insurance.

Portrait of Luis Chavez-Mattos
Edited by Luis Chavez-Mattos, Director, Product··RSS

For Security, Executives and Managers

Extreme close-up of a fanned stack of blank grey cards with one small crimson corner tab sharp against a soft dark blur.

What are tabletop exercise scenarios for incident response?

A tabletop exercise scenario is a short, realistic story about something going wrong (ransomware, a stolen laptop, a breached vendor) that a group talks through, step by step, to test their incident response plan. Nobody touches a system. A facilitator reads the situation aloud, adds new developments called "injects" at set times, and asks the group what they would do, who they would call, and who decides. The output is a list of gaps in the plan, with owners and dates to fix them.

This page gives you the template, ten scenarios to start from, one fully filled-in example, and the AI prompts that write and run the session with you.

TL;DR

  • A good scenario has one trigger, six timed injects, and two or three hard decisions. Everything else is detail.
  • The template below is ready to copy. Fill in your company's systems, people and real contact paths, and you have a session.
  • Run a different scenario each time. Auditors and insurers like to see the plan tested against varied situations, and repeating ransomware every quarter teaches you less.
  • AI is good at drafting scenarios, generating injects and playing the skeptical participant. It is not a source of legal advice, regulatory deadlines or facts about your own environment.
  • The record matters as much as the session. Keep the attendee list, scenario, decisions, gaps and tracked fixes. That is what you show when someone asks for proof you test your plan.
  • Never paste real incident data, customer data or credentials into an AI tool your company has not approved.

Why do teams have to run tabletop exercises?

Plenty of teams run them because someone asked. Security frameworks and compliance programs commonly expect an incident response plan to be tested, and cyber insurance applications often ask whether you do it and how recently. Customers doing vendor reviews ask the same question. NIST's incident response guidance and CISA's free tabletop exercise packages both treat exercises as a normal part of readiness. Check the exact wording in your own audit scope or policy, because requirements differ.

The practical reason is better than the compliance one. Most plans fail at the seams, not in the technical steps. Nobody knows who can authorize taking the website offline. Legal hears about the incident six hours late. The person with the vendor contract is on leave. A conversation around a table finds these problems on a quiet Tuesday instead of during a real event.

What does a good tabletop scenario contain?

Every scenario that works has the same parts:

  1. A trigger. The first piece of evidence anyone sees, written the way it would actually arrive: an alert, a help desk ticket, a call from a customer.
  2. Context. The day, the time, who is on call, what business event is happening (quarter close, a product launch, a holiday weekend). Bad timing is what makes a scenario feel real.
  3. Injects. New information released on a schedule. Each one should force a choice, not just add detail.
  4. Decision points. The two or three moments where leaders must pick between bad options, such as taking a revenue system offline or telling customers before you have all the facts.
  5. Roles. Who has to be in the room: security, IT, legal, communications, HR, a business owner, and an executive with authority to spend money and accept risk.
  6. Objectives. What you are testing, in plain words. "Can we reach legal and the insurer within an hour?" is a good objective. "Test our readiness" is not.
  7. Capture. A way to record what was decided, what was unknown, and what needs fixing.

What is the tabletop exercise template?

Copy this into a document. Brackets are the parts you fill in.

TABLETOP EXERCISE: [Scenario name]

1. LOGISTICS
Date and time: [ ]
Length: [90 minutes recommended]
Facilitator: [name]
Scribe: [name]
Participants (name, role): [ ]
Observers (audit, insurer, board, if any): [ ]

2. OBJECTIVES (pick 3 or fewer)
- Test: [e.g., how fast we reach legal and our insurer]
- Test: [e.g., who can authorize shutting down a system]
- Test: [e.g., whether comms has an approved holding statement]

3. GROUND RULES
- This is a discussion, not a performance review.
- Answer as your role would today, not as you wish it worked.
- If you don't know, say "I don't know" and we log it.
- No real systems are touched.

4. SCENARIO BACKGROUND (read aloud)
Company context: [industry, size, key systems]
Date, time, and business situation: [e.g., Friday 4:45 pm, last day of quarter]
Who is on call: [ ]

5. INJECTS
Inject 1 (T+0): [trigger, as it would arrive]
  Ask: Who sees this first? What do they do? Who do they tell?
Inject 2 (T+15 min): [new evidence or complication]
  Ask: [ ]
Inject 3 (T+30 min): [business impact or outside pressure]
  Ask: [ ]
Inject 4 (T+60 min): [legal, customer, press, or regulator angle]
  Ask: [ ]
Inject 5 (T+4 hours): [escalation or twist]
  Ask: [ ]
Inject 6 (T+24 hours): [recovery and communication question]
  Ask: [ ]

6. DECISION POINTS (the hard calls)
D1: [e.g., Take the order system offline during peak sales?]
D2: [e.g., Notify customers before the investigation is complete?]
D3: [e.g., Engage outside incident response firm and insurer now?]
For each: Who decides? Who must be consulted? What do we need to know first?

7. DISCUSSION QUESTIONS FOR THE END
- Where did we disagree about who owns a decision?
- What information did we not have that we needed?
- Which contact, document, or tool could we not find?
- What would we change in the plan?

8. FINDINGS LOG
| # | Gap or finding | Plan section affected | Owner | Due date | Status |

9. SIGN-OFF
Facilitator: [ ]  Date: [ ]
Executive sponsor: [ ]  Date: [ ]

Which tabletop exercise scenarios should you run?

Rotate through these. Each one stresses a different part of your plan.

ScenarioFirst injectThe hard decisionMust be in the room
Ransomware on a Friday eveningSeveral file servers show a ransom noteIsolate systems now at the cost of business operations, and who is told to consider paymentSecurity, IT, legal, comms, executive
Business email compromiseA finance lead says the CFO asked for an urgent wire, and the CFO denies itRecall the payment, involve the bank, treat the mailbox as breachedFinance, security, legal, executive
Lost laptop with customer dataAn employee reports a laptop stolen from a carIs the disk encrypted, and does this trigger notification dutiesIT, security, legal, privacy, HR
Critical vendor breachedA SaaS supplier emails that "some customer data may have been accessed"Cut the integration, what to tell your own customers, contract rightsProcurement, security, legal, business owner
Departing employee takes dataAn alert shows large downloads from a resignation-notice employeeConfront, suspend access, or monitor quietlyHR, legal, security, manager
Exposed cloud storageA researcher emails that a storage bucket is publicly readableClose it immediately, find out who accessed it, disclose or notCloud engineering, security, legal, comms
Deepfake voice requestA finance team member got a call that sounded like a senior leader asking for a transferVerification rules, who can override themFinance, security, executive
Employee pastes data into a public AI toolA manager discovers a spreadsheet of client data was pasted into a consumer chatbotIs it a reportable incident, how to contain it, what the policy saysSecurity, legal, HR, privacy, the business unit
Extortion email plus press inquiryAn email claims your data was stolen and a reporter asks for commentRespond to either without confirming facts you don't haveComms, legal, security, executive
Major outage during peak eventThe main platform is down on launch day, cause unknownIs it an attack or a failure, and when do you declare an incidentIT, security, operations, comms

Start with the scenario closest to your real risk. Then rotate. A retailer should rehearse a payment-system compromise. A professional services firm should rehearse a client data leak.

What does a filled-in example look like?

Here is the template completed for the first scenario, ransomware on a Friday evening, for a mid-sized company with a customer portal.

Objectives

  • Test how fast we reach legal, the insurer and an outside response firm.
  • Test who can authorize isolating systems during a revenue-critical period.
  • Test whether comms has a holding statement approved in advance.

Background (read aloud) "It is Friday at 4:45 pm, the last business day of the quarter. Sales is closing deals in the CRM. The security lead is on a flight and unreachable for three hours. The IT on-call engineer is a junior team member."

Injects

  • T+0. The help desk gets three tickets: files on a shared drive have strange extensions and a text file named "READ ME" is in every folder. Ask: Who does the help desk call? Is that written down? Does anyone have authority to act without the security lead?
  • T+15 minutes. Another file server shows the same symptoms. The endpoint tool flags an unfamiliar process on a domain controller. Ask: Do we isolate the network segment now? Who approves that? What breaks if we do?
  • T+30 minutes. The VP of Sales learns the CRM is slow and demands to know when deals can close. A customer emails that the portal is returning errors. Decision 1: Do we take the portal offline to contain the spread?
  • T+60 minutes. The ransom note names a cryptocurrency amount and a 72-hour deadline. It claims data was copied. Ask: Who contacts legal? Who finds the insurance policy and its notice requirement? Who has the incident hotline number?
  • T+4 hours. A reporter emails asking whether the company has suffered a breach. An employee has already posted on social media that "everything is down." Decision 2: What do we say, who says it, and what do we refuse to confirm?
  • T+24 hours. Backups exist, but nobody has tested a restore of the file servers in over a year. The CEO asks whether paying is an option. Decision 3: Who advises on payment, what are the legal checks, and what is the recovery plan if we don't?

Typical findings from a session like this

  • The on-call list had a departed employee's mobile number.
  • Nobody could find the cyber insurance policy or its notice terms.
  • The holding statement existed but was never approved by legal.
  • Authority to isolate systems was assumed, not written down.

Each of those goes into the findings log with an owner and a date.

How do you use AI to write and run a tabletop exercise?

AI is useful at three points: drafting the scenario, playing facilitator during the session, and turning messy notes into a report. Use a tool your company has approved, and keep real incident details, credentials and customer data out of it. Describe your environment in general terms.

Prompt 1: Write the scenario

You are an experienced incident response exercise designer. Write a tabletop exercise scenario using the template below.

Our company: [industry, rough size, countries]
Key systems (generic descriptions only): [e.g., customer portal, CRM, payroll system, cloud hosting]
Scenario type: [e.g., business email compromise]
Audience: [security, IT, legal, comms, finance, one executive]
Length: 90 minutes
What we want to test: [3 objectives]
Make it realistic for a Friday afternoon at quarter end.

Produce: background to read aloud, 6 timed injects (T+0, T+15 min, T+30 min, T+60 min, T+4 hours, T+24 hours), each with 2 or 3 discussion questions, 3 hard decision points with the trade-offs on each side, and a list of the roles that must attend. Do not invent specific laws or deadlines. Where a legal or regulatory question comes up, phrase it as "Who confirms what notification obligations apply?" Do not use real company names.

Prompt 2: Act as a live facilitator

Act as the facilitator of a tabletop exercise. I will paste the scenario. You release one inject at a time and wait for me to type what the group decided before you continue.

After each answer from me:
1. Point out any decision that was skipped or assumed.
2. Name any role that was never mentioned but would be involved.
3. Add one realistic complication that follows from what they decided.
4. Then release the next inject.

Be skeptical but not hostile. Do not tell the group the "right" answer. Keep a running list of gaps you notice, and give it to me when I type END.

Scenario: [paste]

Prompt 3: Turn notes into the after-action report

Here are raw notes from a tabletop exercise. Turn them into:
1. A one-paragraph summary of what we tested and how it went.
2. A findings table with columns: finding, why it matters, suggested fix, suggested owner role, priority (high, medium, low).
3. Three questions we left unanswered.
4. A list of anything in the notes that sounds like a guess rather than a decision, so we can verify it.

Use only what is in the notes. If something is unclear, flag it instead of filling it in.

Notes: [paste]

Prompt 4: Make the next one harder

Here is our last scenario and its findings: [paste]. Write a new scenario that tests a different failure mode, and add one inject that checks whether the fix from our last finding actually works under pressure.

Where should you not trust AI?

Check these every time:

  • Laws and deadlines. Notification windows vary by jurisdiction, industry and contract. AI will sometimes state them confidently and wrongly. Have legal supply them.
  • Your environment. AI does not know which of your systems are critical, who actually has authority, or what your insurance policy says. Any of those it states is a guess.
  • Ransom payment advice. This involves legal, sanctions and insurer considerations. Treat anything AI says as a prompt for a question to counsel, not guidance.
  • Realism. Generated injects can be too tidy. Real incidents are ambiguous, and information arrives contradictory and late. Edit scenarios to add noise and uncertainty.
  • Confidentiality. Use generic descriptions and an approved tool. A tabletop about a breach should not create one.

How do you run the session itself?

  1. Pick the scenario and objectives two weeks ahead. Share only the objectives, not the scenario details.
  2. Invite decision makers, not just responders. The most useful findings come from executives and legal discovering what they would be asked to decide.
  3. Open for five minutes. Read the ground rules. Say clearly that gaps are the point.
  4. Read the background, then release injects one at a time. Pause after each for discussion. Resist the urge to answer your own questions.
  5. Ask "who decides?" often. Most plans are vague here.
  6. Ask "how do you know?" for every assumption. If it rests on a document, ask someone to find the document in the room.
  7. Keep the scribe separate from the facilitator. One person cannot do both well.
  8. Spend the last 15 minutes on findings. Assign an owner and a date to each before anyone leaves.
  9. Send the report within a week. A late report is a forgotten one.

What should you keep for auditors and insurers?

Keep a single folder per exercise with the scenario, the attendee list with roles, the date, the decisions recorded, the findings log, and evidence that fixes were completed. If an auditor or an insurer asks whether you test your plan, you want to show the last exercise and the proof that the findings were closed, not only that a meeting took place. Also update the incident response plan itself after each session, and note the revision date.

Can you build a findings tracker instead of using spreadsheets?

The part of tabletops that usually breaks is follow-through: findings sit in a document and nobody chases them. With Remy, you could describe a small internal app in plain English: a scenario library, a schedule of upcoming exercises, a findings log with owners, due dates and status, sign-in with role-based access, and a view that shows overdue fixes. Remy builds and deploys it as a full-stack app you own, so audit evidence lives in one place. Describe the fields from the template above and you have the spec.

Frequently Asked Questions

5 questions

How often should we run a tabletop exercise?

Most teams run at least one a year, and many aim for quarterly or twice a year with different scenarios. Check your audit scope, framework or insurance policy for any stated frequency. Also run one after a major change, such as a new cloud provider, a merger or a significant incident.

How long should a tabletop exercise take?

Plan 90 minutes to two hours for a first session, including the debrief. Shorter 45 to 60 minute versions work well for a single team, such as a finance-only session on wire fraud. If the session runs long, you probably have too many injects.

Who should facilitate?

Someone who is not responsible for the plan being tested, since that person tends to defend it. A security colleague from another team, internal audit, or an outside facilitator all work. An AI facilitator can support the human by tracking gaps and generating injects, but a person should run the room.

Is a tabletop exercise the same as a simulation or red team test?

No. A tabletop is a discussion about decisions and communication, with no live systems involved. A technical simulation or red team test attacks real systems to test defenses and detection. Tabletops test whether people and process work. Many programs use both.

Do we need to include executives?

Yes. Executives approve spending, accept risk and speak to customers, boards and regulators. If they have never walked through a scenario, the first time they face those decisions will be during a real incident.

How we source, write, and correct our articles. Editorial standards

Related articles

Straight to your inbox.
New pieces as they publish. No noise.