CISO Cyber Incident Response Plan: Prepare, Test & Respond

A cyber incident response plan is the document that pre-decides what nobody can decide well at 02:00: who owns which call, what triggers activation, and when each notification clock starts running. This guide covers what the document must specify, the clocks it must encode, one worked example, and how to test the wording before an incident exposes its weaknesses.

Layered geometric pathways through protective structures, representing cyber incident response planning.

Key Takeaways

  • Every clause in a cyber incident response plan represents a decision made in advance. A section that decides nothing remains a serious failure point.

  • Regulators influence the plan’s contents. NYDFS §500.16(a)(1) identifies nine required areas, including roles and levels of decision-making authority.

  • Notification clocks begin with an internal event or determination. The plan must name who makes that determination and on what evidence.

  • A cloud annex defines provider support and evidence access. A ransomware annex defines payment authority and notification duties, not negotiation tactics.

  • A review confirms that the document exists. An exercise with a failure condition shows whether its wording produces timely decisions.

A cyber incident response plan pre-decides how an organization detects, classifies, contains, and reports an incident. It names each decision owner, defines what activates the plan, and identifies when external notification clocks begin.

Most plans pass their annual review. Fewer have been tested against a difficult night. A review confirms that the document exists and covers the required topics. It cannot show whether those clauses produce timely decisions under pressure.

Many CISOs inherit a plan written for an auditor and stored in a shared drive. Its most important clauses may never have been tested against a running clock. This guide explains what the document must specify, which deadlines it must encode, and how to test whether its wording holds.

What Is a Cyber Incident Response Plan?

An incident response plan tells an organization what to do when a cybersecurity event is suspected or confirmed. It defines what counts as an incident and names the owner of each decision. It also establishes activation, escalation, and external-notification conditions.

Practitioners often describe the plan as a runbook. Regulators treat it as a required artifact with specified contents.

New York's Department of Financial Services is the most specific about contents. 23 NYCRR §500.16(a)(1) requires covered entities to establish a written plan and lists nine areas it must address. Among them: the goals of the plan, internal response processes, external and internal communications, remediation requirements, documentation and reporting, recovery from backups, and root cause analysis. One more sits at the center of this article: "the definition of clear roles, responsibilities and levels of decision-making authority." That is a table of contents drafted by a regulator.

The plan also documents a lifecycle, the sequence a team follows from detection through recovery. The revised incident handling guidance moves the improvement work that exercises produce out of a terminal preparation phase and into a category that feeds every other function continuously. Those stages have their own article, Incident Response Process: 7 Critical Steps for Security Teams. This page stays on the document.

One framing matters more than the section count. Every clause is a decision made in advance, in daylight, by people who had time to think. Reflex Security runs exercises against these documents, and the recurring finding is rarely a missing section. It is a section that decides nothing.

The Cyber Incident Response Plan Steps Every CISO Should Define

The incident response plan steps a CISO has to define are document sections, not process stages. A complete plan specifies nine things:

  • Scope and activation criteria: what conditions turn the plan on, and who turns it on.

  • Severity classification, with worked examples rather than adjectives.

  • Named owners and named delegates for every decision the plan reserves.

  • Notification triggers and the clocks they start.

  • Evidence handling and legal-hold instructions.

  • External-party contacts, with the contracts that make them answerable.

  • The communications approval path, including who speaks and who signs off.

  • Recovery and stand-down criteria.

  • The review and test cycle that keeps the other eight true.

Nine sections is a checklist. Determining the exact moment each notification clock starts is the hard part.

Encode the Notification Clocks

A notification deadline never starts when the attacker arrives. It starts on an event the plan itself has to define. That event is almost always an internal judgment call.

Obligation

What starts the clock

Who the plan must name

What the document has to contain

GDPR Article 33: notice to the supervisory authority

Becoming "aware" of a personal data breach. Notice follows without undue delay and, where feasible, not later than 72 hours; later notice must be accompanied by reasons for the delay.

The person who determines awareness, and the evidence standard they apply

A written awareness test, a named determiner, and a drafted holding notice for the phased submission Article 33(4) permits

NYDFS 23 NYCRR §500.17(a): notice to the Superintendent

Determining that a cybersecurity incident has occurred at the entity, its affiliates, or a third-party service provider. Notice is due as promptly as possible and no later than 72 hours.

The determiner, plus the out-of-hours path that reaches them

Third-party incident intake, a named owner for the DFS submission, and the continuing duty to update the Superintendent

US state breach-notification laws: a variance row, not a deadline row

Different events in different states. Colorado runs from "determination that a security breach occurred." California, as amended effective January 1, 2026, runs from discovery or notification of the breach.

Whoever maintains the jurisdiction map and decides which residents are in scope

A jurisdiction map keyed to residency, attorney-general thresholds, and the separate deadlines those thresholds carry

The first two obligations both set a 72-hour limit, but their clocks start differently. Article 33 begins with awareness, while Part 500 begins with a determination. Colorado and California both allow 30 days, and they begin counting at different moments. The plan therefore cannot encode one universal deadline. It must encode a determination process, a named determiner, and a map of required recipients.

Article 34 then adds a second, separately triggered duty to the same document: communication to affected individuals where the breach is likely to result in a high risk to their rights and freedoms. A public company carries one more clock, the SEC's 2023 cyber disclosure rule, which attaches to a materiality determination rather than to containment; the board and disclosure layer belongs to The C-Suite Cybersecurity Tabletop Exercise Playbook.

Add Cloud and Ransomware Annexes

Two annexes finish the document. A cloud incident response plan annex records what the provider will supply under shared responsibility. It should identify available log sources, retention periods, the support path for forensic requests, and the person authorized to open one. The organization should document that path before an incident because it is difficult to discover at speed.

A ransomware incident response plan annex records decision rights and clocks, not negotiation strategy. It identifies who authorizes and signs an extortion payment. It must also account for NYDFS §500.17(c), where applicable. That provision requires notice within 24 hours and a written explanation within 30 days. Payment strategy itself belongs to Ransomware Tabletop Exercise: A Healthcare Crisis Scenario.

Who Owns What? Roles and Decisions During a Cyber Incident

The plan's job is not to describe the response team. Its job is to encode ownership so precisely that nobody negotiates it during a weekend turn. Part 500 asks for "clear roles, responsibilities and levels of decision-making authority" for a reason. Roles are the easy half. Levels of authority are where documents go quiet.

Five clause types carry that weight:

  • The owner clause, naming a role and the current person in it, with the date the name was last verified.

  • The delegate clause, naming who acts when the owner is unreachable, and stating how the delegate learns they are on.

  • The authority-to-act clause, stating what the owner may do without further approval, including limits on spend and on deliberate service disruption.

  • The escalation clause, stating the condition that transfers ownership upward, in terms a night-shift operator can evaluate without a phone call.

  • The determination clause, naming who declares that an incident has occurred and on what evidence, because that declaration starts the regulatory clocks.

The full cast of an incident response team, and what each function does once the plan is activated, belongs to Incident Response Roles and Responsibilities: Who Owns What? What belongs here is the drafting failure. A plan that leaves ownership implicit reads well in review, because implicit ownership looks like consensus on the page. It only fails once.

Cyber Incident Response Plan Example: From Detection to Recovery

The most useful incident response plan example tests the organization’s own document against a night it was not written for.

Consider a Colorado community water system serving approximately 40,000 residents. Its environment includes SCADA, chemical-dosing HMIs, and a vendor remote-access path for pump controls. Three people manage IT without a security operations center. A customer billing database holds names, addresses, and payment details.

The Incident at 02:14

At 02:14, the operator sees a dosing setpoint she did not enter. The historian shows the vendor’s remote-access account authenticating at 20:09 the previous evening.

The plant can maintain finished-water quality through manual dosing for approximately six hours. After that, the superintendent must decide what happens to the distribution system.

Who authorizes the switch to manual dosing? Who tells the state primacy agency? The plan answers neither question. Three clauses explain why.

Three Clauses That Fail

  1. “The utility will notify the relevant authorities as appropriate.”
    Every key word defers a decision. The clause names no authority, trigger, owner, or clock. Its author assumed the reader would know what to do. At 02:14, that reader is a shift operator who has never opened the file.

  2. “The IT Manager serves as Incident Commander.”
    The IT manager resigned in March. The plan’s revision page records a June review, complete with initials and a date. The review confirmed that the section existed, but nobody called the listed number.

  3. “Escalate to executive leadership if the incident is deemed significant.”
    The clause does not say who makes that determination, what evidence they need, or how quickly they must decide. Passive language conceals the missing owner, while the vague threshold allows the clause to survive review.

The Clocks the Plan Missed

The billing database could bring the utility under Colorado’s breach-notification regime. The applicable statute depends on how the utility is organized.

A municipal system answers to C.R.S. §24-73-103, which covers districts, instrumentalities, and political subdivisions. An investor-owned system answers to §6-1-716.

Both statutes allow 30 days after determining that a breach occurred. Neither names the organization’s determiner, so the plan must do so.

The drinking-water system carries a separate obligation. Under the EPA’s Public Notification Rule, a Tier 1 situation requires notification to people who may drink the water within 24 hours. The two duties have different triggers, deadlines, and audiences. The existing plan encodes neither.

The utility also maintains an emergency response plan certified under AWIA section 2013. AWIA requires community water systems serving more than 3,300 people to address cybersecurity and review the underlying assessment at least every five years. Yet neither document identifies who owns the dosing decision at 02:14.

A third clock may be coming. CISA’s proposed CIRCIA rule could apply a 72-hour reporting duty to qualifying water systems. As of September 24, 2026, CISA had not issued the final rule. Plans should label the obligation as pending and confirm its status during each revision.

Managing the escalating event is a separate discipline, covered in Incident Response Management: From Detection to Crisis Recovery. The document-level verdict is narrower: the plan passed review but answered none of the operational or notification questions raised at 02:14.

Why a Cyber Incident Response Plan Needs to Be Tested

Testing begins with a basic question: what is an incident response plan expected to deliver on its own? It gives a frightened team a shared reference at the worst hour of their year and satisfies a real regulatory obligation. The document also provides one place to make decisions calmly, before anyone is tired, angry, or on a call with a reporter. A good plan is worth writing and worth keeping current.

It is still not evidence of capability. A review tests whether the document exists, whether the required sections are present, and whether someone signed the revision page. It does not test whether the named owner will be a bottleneck dealing with too many responsibilities. It does not test whether the delegate knows they are the delegate. Nobody audits whether a shift operator can start a 72-hour clock without a lawyer in the room.

Part 500 anticipates the gap. §500.16(d) requires covered entities to test the incident response plan periodically, at minimum annually, "with all staff and management critical to the response," and to revise it as necessary. The same regulation that mandates the document mandates the rehearsal. Annual testing duties appear across the common assurance frameworks too, and who is required to test reaches further than most CISOs assume.

One catch undercuts the mandate. Tabletop exercises capture observations of a hypothetical conversation rather than operational practice, which is why exercises with purely theoretical consequences produce agreement where nobody fails rather than evidence. An assessment where nothing is actually executed says nothing about a document that has never been contradicted. 

Ways to Test a Cyber Incident Response Plan: Tabletops, Simulations & Drills

Three formats produce three kinds of evidence about a plan. A tabletop shows what the document says and does not say. A drill shows whether one named mechanism works. A simulation shows whether the wording produces a decision. The required evidence should determine the format.

A tabletop deserves consideration first because it is the least expensive effective option. Walking a group through a scenario can expose stale names, vague notification clauses, and unclaimed decisions within two hours. For finding what a document is missing, a discussion beats a simulation on both cost and speed. That advantage is real.

What a discussion cannot do is apply pressure. Participants describe what they would do, the group converges, and "as appropriate" survives untested because nobody was forced to act on it. A drill narrows the aperture instead: one call tree, one restore, one submission through the regulator's portal, timed and repeatable. Useful, and deliberately small.

Four requirements decide which format the plan needs, and each one names the evidence before it names the instrument. Reflex is built for the third kind.

Test the Named Owner and Delegate

Proof the named owner and the named delegate were reachable and acted. That evidence exists only if the people named in the plan work the incident themselves, with any stand-in recorded as a stand-in. Reflex runs it as a live simulation, with AI agents playing the adversary and stakeholders who push back in real time. This is what separates simulating a live crisis from describing one.

Measure When the Clock Starts

Proof the clock started on time and was started by the person the plan names. This requires timestamps against a scenario that does not pause. Reflex escalates when the team stalls, so the interval between the triggering event and the determination is measured rather than assumed. Participants work the incident through a web-based Investigation Console. Every finding in the report traces back to the moment in the simulation that produced it.

Use a Scenario Based on the Environment

Proof the scenario resembles the environment the plan governs. Generic injects test a generic document. Reflex builds from the organization’s domain. Adaptive scenario generation produces five customized scenarios in 10 to 15 minutes using its systems, vendors, and people.

Produce Evidence for Review

An artifact that can be given to an auditor, board, or insurer. A Reflex exercise runs 60 to 120 minutes and closes with an After Action Report that typically exceeds 54 pages. It records decisions, timings, and the clauses that failed to produce a decision. The report supports each finding with evidence from the exercise rather than presenting a standalone score.

Reflex tests whether a current plan produces timely decisions under pressure. Bring your document and a scenario tied to your environment, then examine which clauses fail to guide the team.

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 long should a cyber incident response plan be?

Short enough that a shift operator can find an answer in two minutes. Most usable plans run 15 to 30 pages, with the decision tables near the front and the detail in annexes. A 120-page plan is usually a policy, a runbook and a contact list that nobody separated.

Should it be one document or a set of playbooks?

One plan, several annexes. The plan carries activation, severity, ownership, authority and the notification clocks, because those are identical whatever happened. Scenario-specific detail belongs in annexes: ransomware, cloud provider, third-party compromise. Duplicating the authority clauses into each playbook guarantees three versions and one argument.

Where should the plan live so it is reachable when systems are down?

In at least two places, one of which does not depend on your identity provider or your corporate network. Printed copies at the sites that need them, plus a copy the response team can reach from a personal device. A plan stored inside the system under attack is not a plan.

Does the plan need board or executive approval?

Some regimes expect it, and all of them benefit from it, because the clauses that matter reserve authority the board ultimately delegates. What approval actually buys you is an executive who has read the escalation clause before the night it fires. Check your own regulator's wording for whether approval is required or merely wise.

Is a downloadable template a safe starting point?

As a structure, yes. As a plan, no. Templates supply the section headings and the language that defers decisions: as appropriate, relevant authorities, if deemed significant. Keep the headings, then replace every deferring phrase with a name, a threshold or a clock. The editing is the work.

How often does the plan need revising beyond the annual test?

Whenever a name in it changes, which in practice is more often than annually. Owners, delegates, vendor contacts and the out-of-hours path age fastest. Date-stamp each name when it was last verified, and treat an unverified name as a gap rather than as content.

References

Cybersecurity and Infrastructure Security Agency. (n.d.). Cyber Incident Reporting for Critical Infrastructure Act of 2022 (CIRCIA). Cybersecurity and Infrastructure Security Agency. https://www.cisa.gov/topics/cyber-threats-and-advisories/information-sharing/cyber-incident-reporting-critical-infrastructure-act-2022-circia

Under CISA’s proposed rule, covered entities would report covered cyber incidents within 72 hours of reasonably believing an incident occurred. They would report ransom payments within 24 hours. As of September 24, 2026, CISA had not issued the final rule. The requirements therefore remain pending and should be checked again before publication.

FindLaw. (n.d.). Colorado Revised Statutes § 6-1-716: Notification of security breach. FindLaw. https://codes.findlaw.com/co/title-6-consumer-and-commercial-affairs/co-rev-st-sect-6-1-716/

Applies to a covered entity handling personal information in the course of a business, vocation, or occupation. The 30-day period runs from determination that a security breach occurred. Attorney-general notice threshold: 500 Colorado residents.

FindLaw. (n.d.). Colorado Revised Statutes § 24-73-103: Notification of security breach by governmental entities. FindLaw. https://codes.findlaw.com/co/title-24-government-state/co-rev-st-sect-24-73-103/

The governmental-entity counterpart, reaching the state, counties, incorporated cities and towns, authorities, and every other kind of district, instrumentality, or political subdivision. A municipal water system falls here rather than under § 6-1-716. Same 30-day clock from determination and same 500-resident attorney-general threshold.

Intersoft Consulting. (n.d.). Art. 33 GDPR: Notification of a personal data breach to the supervisory authority. General Data Protection Regulation (GDPR). https://gdpr-info.eu/art-33-gdpr/

State the condition, not a flat deadline. Notice runs from becoming aware of a personal data breach, and is due without undue delay and, where feasible, not later than 72 hours after awareness. Where notice is later than that, it must be accompanied by reasons for the delay. Article 33(4) permits phased submission.

Intersoft Consulting. (n.d.). Art. 34 GDPR: Communication of a personal data breach to the data subject. General Data Protection Regulation (GDPR). https://gdpr-info.eu/art-34-gdpr/

A separately triggered duty, owed where the breach is likely to result in a high risk to the rights and freedoms of natural persons.

Legal Information Institute. (n.d.). 23 NYCRR § 500.16: Incident response and business continuity management. Cornell Law School Legal Information Institute. https://www.law.cornell.edu/regulations/new-york/23-NYCRR-500.16

Verified September 10, 2026. §500.16(a)(1) lists the nine required areas, including "the definition of clear roles, responsibilities and levels of decision-making authority." §500.16(d) requires testing "periodically, but at a minimum annually… with all staff and management critical to the response." The 2023 Second Amendment added "and business continuity management" to the section title.

Legal Information Institute. (n.d.). 23 NYCRR § 500.17: Notices to superintendent. Cornell Law School Legal Information Institute. https://www.law.cornell.edu/regulations/new-york/23-NYCRR-500.17

Verified September 10, 2026. §500.17(a) sets notice to the superintendent no later than 72 hours after determining that a cybersecurity incident has occurred at the covered entity, its affiliates, or a third-party service provider. §500.17(c) requires notice of an extortion payment within 24 hours and a written description of why it was necessary within 30 days.

US Securities and Exchange Commission. (2023, July 26). SEC adopts rules on cybersecurity risk management, strategy, governance, and incident disclosure by public companies (Press release 2023-139). US Securities and Exchange Commission. https://www.sec.gov/newsroom/press-releases/2023-139

An Item 1.05 Form 8-K is "generally be due four business days after a registrant determines that a cybersecurity incident is material." The clock attaches to the materiality determination, not to containment.

State of California. (n.d.). California Civil Code § 1798.82. California Legislative Information. https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=CIV&sectionNum=1798.82

Verified September 10, 2026. As amended by SB 446, effective January 1, 2026, § 1798.82(a)(2)(A) requires disclosure "within 30 calendar days of discovery or notification of the data breach."

US Environmental Protection Agency. (n.d.). Public notification rule. US Environmental Protection Agency. https://www.epa.gov/dwreginfo/public-notification-rule

A Tier 1 situation obliges the system to notify people who may drink the water within 24 hours.

US Environmental Protection Agency. (n.d.). AWIA section 2013. US Environmental Protection Agency. https://www.epa.gov/waterresilience/awia-section-2013

Community water systems serving more than 3,300 people must address cybersecurity in the certified emergency response plan and review the underlying risk assessment at least every five years.

All URLs verified September 10, 2026. No tracking parameters.

Cite this article

Cassio Goldschmidt.

CISO Cyber Incident Response Plan: Prepare, Test & Respond.

Reflex Security.

https://reflexsecurity.com/guides/cyber-incident-response-plan

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.