Implementation

Incident response plan: template and phases

· 13 min read ·

An incident response plan is the document that sets out, in advance, who does what when the organisation's security is compromised. It turns a confused situation into a sequence of predictable steps: detect, contain, eradicate, recover and notify whoever the law requires. This article offers a practical template, with NIST's classic phases and what changed in the 2025 revision, the team's roles, the notification deadlines owed to Brazil's data protection agency, Central Bank and insurance supervisor, and a method for testing the plan before it is needed for real.

The urgency is real. In the 2025 edition of the Cost of a Data Breach report, organisations took a mean of 241 days to identify and contain a breach, the shortest time in nine years (IBM, Cost of a Data Breach Report 2025). Even at the best result in recent years, that is almost eight months between compromise and containment. A rehearsed response plan is one of the most direct ways to shorten that window.

What does an effective incident response plan look like?

An effective plan organises the response around three pillars: process, people and evidence. The process follows clear phases. People have roles assigned before the crisis. Evidence is collected in a way that supports both the technical investigation and accountability to regulators. Without all three, the plan becomes a document that sits on a shelf.

Security team monitoring systems in an operations centre while responding to an incident

The document does not need to be long, but it must answer hard questions under pressure. Who has the authority to shut down a production system? Who do you call at two in the morning on a Sunday? How do you preserve the integrity and provenance of a compromised disk? NIST itself lists what most incident response policies have in common: a statement of management commitment, purpose, scope, definitions of events and incidents, roles and authorities (for example, who may disconnect or shut down an asset), criteria for prioritising incidents and estimating severity, and performance measures (NIST SP 800-61 Rev. 3, section 2.3).

In practice, a good template covers scope, severity classification criteria, a call tree, playbooks by threat type and the triggers for legal notification. NIST recommends documenting procedures at least for the most common types of incident. Playbooks should follow the most frequent entry points: in the 2026 Data Breach Investigations Report (DBIR), Verizon reports that 31% of breaches began with the exploitation of vulnerabilities, which overtook stolen credentials as the main way in (Verizon, 2026 DBIR). Deciding which scenarios get a playbook first is a prioritisation exercise, much like building a risk matrix.

What are the phases of incident response?

The most widely cited model comes from the NIST SP 800-61 guide. In Rev. 2, from 2012, the cycle had four phases: preparation; detection and analysis; containment, eradication and recovery; and post-incident activity. Many plans split the third phase into three and treat lessons learned as a stage of its own, which produces the five- or six-step design common in the market. Preparation is the only phase that happens before the incident, and it has the greatest influence on how the others turn out.

In April 2025, Rev. 3 superseded Rev. 2 and dropped the circular cycle. NIST explains that in 2012 incidents were rarer, narrow in scope and usually resolved within a day or two. Today they are frequent and recovery can take weeks or months. The new version organises the response around the six functions of the NIST CSF 2.0 and maps them to the old phases:

Phase of the previous model (Rev. 2)CSF 2.0 functions (Rev. 3)
PreparationGovern, Identify (all categories) and Protect
Detection and analysisDetect and Identify (Improvement category)
Containment, eradication and recoveryRespond, Recover and Identify (Improvement category)
Post-incident activityIdentify (Improvement category)

NIST says each organisation should use the life cycle model that suits it best. The classic phases remain useful for writing the plan. What changes with Rev. 3 is the emphasis: preparation becomes part of ongoing risk management, and lessons learned should be shared as soon as they are identified, not only once recovery is over.

In preparation, the organisation builds the team, defines playbooks, secures access to forensic tools and trains people. In detection and analysis, alerts are triaged and the incident is confirmed, categorised and prioritised. The sooner this phase ends, the smaller the damage tends to be.

Containment isolates the problem to stop it spreading, often in two steps: short-term containment, which stems the damage, and long-term containment, which prepares the clean-up. Eradication removes the root cause, whether malware, a compromised credential or an exploited vulnerability. Recovery restores systems and monitors them closely to confirm that the attacker has really gone. Rev. 3 recommends checking the integrity of backups and other restoration assets before using them (RC.RP-03). Knowing quickly which assets are exposed depends on good visibility of your cyber exposure.

The threat profile matters when designing playbooks. On its 2026 DBIR page, Verizon states that ransomware is present in 48% of breaches, although payouts are shrinking and businesses frequently choose not to pay (Verizon, 2026 DBIR). A dedicated ransomware playbook, covering isolation and verification of backups, criteria for the payment decision and when to bring in the legal team, deserves priority.

Who does what during a crisis?

During an incident, confusion over roles costs precious hours. That is why the plan should name functions, not people, and map each function to a primary holder and a deputy. The minimum structure includes an incident coordinator, technical analysts, a legal and privacy focal point, communications and a link to executive leadership.

The incident coordinator leads the response, makes containment decisions and keeps the chronological log. Technical analysts investigate, collect evidence and carry out containment and eradication. Legal and the data protection officer (DPO) assess notification obligations and keep decisions defensible. Communications handles internal messages, messages to customers and, when needed, the press. The executive link approves high-impact decisions and informs the board. Rev. 3 describes similar roles: leadership oversees the response, allocates funding and may decide high-impact actions such as shutting down or rebuilding critical services, while asset owners provide recovery priorities.

In Brazil's financial sector, some of these roles are a regulatory requirement. CMN Resolution No. 4,893/2021, issued by the National Monetary Council (CMN) for institutions authorised by the Central Bank of Brazil (BCB), requires a director responsible for the cybersecurity policy and for executing the incident action and response plan (art. 7), and the plan must be approved by the board of directors or, where there is none, by the executive board (art. 9).

Third parties complicate the organisation chart. In the 2026 DBIR, breaches involving a third party reached 48% of the total, up 60% on the previous edition (Verizon, 2026 DBIR). NIST treats response with service providers as a shared responsibility: the contract should define information flows, coordination and the authority to act on the organisation's behalf. The plan should set out how to engage compromised suppliers and require their cooperation, a point that links directly to vendor risk management and, for banks, to the treatment of third parties under CMN Resolution 4,893.

The plan still depends on detection and response tools. What a risk platform such as Futture's Eagle can do is have the context ready before the crisis. The business continuity module connects critical processes, assets, applications, dependencies and RTO/RPO, and the third-party risk module brings together the supplier register, criticality and contracts. With that map in place, the response team can look up which processes and contracts an incident may affect instead of rebuilding it under pressure.

What are the notification obligations to the ANPD and other regulators?

Under the LGPD (Brazil's General Data Protection Law, art. 48), the controller must notify the National Data Protection Agency (ANPD) and data subjects of any security incident that may cause significant risk or harm. ANPD Resolution No. 15/2024 defines when this applies (art. 5): the incident must be capable of significantly affecting data subjects' interests and fundamental rights and, at the same time, involve at least one of these criteria: sensitive data; data of children, adolescents or older people; financial data; system authentication data; data protected by legal, judicial or professional secrecy; or large-scale data.

The deadline is 3 working days from the controller learning that the incident affected personal data, both for the ANPD (art. 6) and for data subjects (art. 9). Notification to the ANPD is made through an electronic form, by the DPO or a duly appointed representative, with the documents proving that relationship or power of representation submitted within the same deadline. The information may be supplemented, with justification, within 20 working days of the notification, according to the ANPD's official guidance. Small processing agents have the deadlines doubled (art. 6, § 8). And every incident, whether notified or not, must be entered in a register kept for at least five years (art. 10).

In practice, the clock starts with "knowledge", so the plan needs to define the moment at which the incident is considered known and who records that date. This is a legal and a technical decision at the same time. The resolution itself preserves deadlines set by specific legislation, and a single incident can start more than one clock:

RecipientRuleWhat to notifyDeadline
ANPDANPD Resolution No. 15/2024, art. 6Incident that may cause significant risk or harm to data subjects3 working days from learning that the incident affected personal data; supplementary information within 20 working days
Data subjectsANPD Resolution No. 15/2024, art. 9The same incident, in plain language3 working days from the same point
Central Bank of BrazilCMN Resolution No. 4,893/2021, art. 20, IIIRelevant incidents and interruptions of relevant services that amount to a crisis, and the steps taken to resume activitiesTimely notice; the resolution sets no deadline in days, and the institution must document its crisis criteria
SUSEP (Brazil's insurance supervisor)SUSEP Circular No. 638/2021, art. 8Relevant incidents, according to parameters set by the supervised entity itself5 working days from becoming aware of the event

The plan should list every applicable regulator and its deadline, so that the ANPD notification does not overshadow the others. Payment institutions are not covered by Resolution 4,893 and follow the Central Bank's own rules instead (art. 1, sole paragraph). The details are in our articles on ANPD regulations, CMN Resolution 5,274/2025, which updated Resolution 4,893, and SUSEP Circular 638.

How do you test and improve the plan with exercises?

A plan that has never been tested is a hypothesis, not a procedure. The cheapest way to validate it is a tabletop exercise: the team walks through a fictional scenario step by step and finds where the process stalls. Outdated contact lists, unclear decision-making authority and missing access to tools show up in these exercises rather than in the real incident.

None of the rules cited sets an exercise frequency that applies to every company. One exercise a year is common practice, with extra sessions after significant changes to infrastructure, staff or regulation. In the financial sector, Resolution 4,893 requires the plan to be reviewed at least once a year (art. 10) and the policy to include guidelines for the incident scenarios used in business continuity tests (art. 3, V, "a"). NIST's Rev. 3 treats exercises and tests, including those run with critical suppliers, as a source of improvement (ID.IM-02) and refers to NIST SP 800-84 for simulations and tabletop exercises.

Useful scenarios include ransomware in production, a leak through a supplier and a compromised privileged credential. If the company answers to more than one regulator, run the scenario with every clock in the same exercise. At the end, each failure becomes an action with an owner and a deadline, and the plan is revised. This continuous improvement cycle is the same one that underpins continuous internal audit.

Measuring matters too. Track time to detect, time to contain and completeness of evidence across exercises, ideally as key risk indicators with a threshold and an owner. If the numbers improve from one exercise to the next, the investment is working. Rev. 3 asks for incident data to be collected with its integrity and provenance preserved (RS.AN-07), and that rigour is what separates a defensible response from a pile of emails gathered after the fact.

Conclusion

Start with what gives the most return for the least effort: write the call tree and book a tabletop exercise for next month. With the team in the room, walk through a ransomware scenario and time how long it takes until someone knows who to call, when the incident became "known" and which deadlines have started running. The gaps that appear in that first hour are exactly the ones that would have been expensive in a real incident. From there, formalise the phases, assign the roles and repeat the test until the numbers improve.

Further reading: ANPD regulations: a map of what your company must comply with; NIST CSF 2.0: the six functions and what Govern changed; SUSEP Circular 638: cybersecurity rules for Brazilian insurers.

Frequently asked questions

What is an incident response plan?

It is the document that defines how the organisation detects, contains, eradicates and recovers from a security incident, and who it must notify. It sets out roles, decision flows and notification obligations before a crisis happens.

What is the deadline for notifying the ANPD of an incident?

Three working days from the controller learning that the incident affected personal data, when it may cause significant risk or harm to data subjects (ANPD Resolution No. 15/2024, arts. 6 and 9). The same deadline applies to data subjects, and the information may be supplemented within 20 working days of the notification.

What are the phases of an incident response plan?

The classic model in NIST SP 800-61 Rev. 2 has four phases: preparation; detection and analysis; containment, eradication and recovery; and post-incident activity. Rev. 3, published in April 2025, replaced that cycle with a model based on the six functions of the NIST CSF 2.0 and says each organisation may use the model that suits it best.

How often should the plan be tested?

None of the rules cited sets a single frequency. One exercise a year is common practice, with extra sessions after significant changes to infrastructure, staff or regulation. Institutions authorised by Brazil's Central Bank must review the plan at least once a year (CMN Resolution No. 4,893/2021, art. 10).

Are there other notification deadlines besides the ANPD?

Yes, depending on the sector. Insurers and other entities supervised by SUSEP must report relevant incidents within five working days (SUSEP Circular No. 638/2021, art. 8). Institutions authorised by Brazil's Central Bank must give timely notice of relevant incidents that amount to a crisis (CMN Resolution No. 4,893/2021, art. 20). These deadlines run in parallel with the ANPD's.

Official references

About the author

Founder, Futture.ai | CISSP

Cybersecurity, governance and risk management specialist with over 25 years of industry experience. CISSP-certified, founder of Network Secure and Futture.ai.

Read next

← Back to the blog