Back to Articles
    Technology & Security

    Breach Notification Day One: An Incident Response Plan for Small Nonprofits

    The morning your nonprofit discovers encrypted files, a hijacked inbox, a missing laptop, or a letter from a vendor saying "your data may have been affected," the decisions you make in the next 24 to 72 hours shape everything that follows: what evidence survives, what your insurance pays for, which legal deadlines you meet, and whether donors and clients still trust you afterward. This guide walks a small US nonprofit through that window step by step. It covers who to call first, how to contain the problem without destroying the evidence, how to figure out what data was involved, how state breach notification laws, the HIPAA Breach Notification Rule, and the FTC Health Breach Notification Rule fit together, and how to tell your board, donors, and clients. It also shows where AI can help you prepare a one-page plan and practice it in advance, where AI must stay out of the response, and what to do when the incident involves AI itself.

    Published: November 10, 2026•19 min read•Technology & Security
    A small nonprofit team reviewing security alerts on a laptop, representing data breach incident response planning

    Most small nonprofits have spent at least some time on prevention. They have turned on multifactor authentication, run a phishing training, maybe bought a cyber insurance policy. Far fewer have thought through what happens on the day prevention fails. That is understandable. Incident response sounds like something for organizations with a security team and a 24-hour operations center, not a twelve-person community organization with a part-time IT contractor. But the smaller you are, the more a plan matters, because nobody on staff has done this before and the first instincts people have under stress (wipe the computer, pay the ransom, email everyone, say nothing) are often the wrong ones.

    The stakes are real for organizations of every size. IBM's Cost of a Data Breach Report 2025 put the average cost of a breach in the United States at a record $10.22 million, driven in part by regulatory penalties and detection costs. A small nonprofit's breach will cost a tiny fraction of that, but it will also land on a tiny fraction of the budget. And nonprofits are not exempt from scrutiny. When the fundraising software vendor Blackbaud suffered a ransomware attack in 2020 that exposed data from thousands of nonprofit customers, the company later agreed to a $49.5 million settlement with state attorneys general, and thousands of nonprofits had to work out their own obligations to the donors whose information was involved. A vendor's breach became every customer's breach notification problem.

    This article is deliberately narrow. It does not try to teach you how to prevent attacks; our guide to strengthening cybersecurity on a small budget covers that. It does not walk through how to buy or evaluate coverage; see our article on cyber insurance for nonprofits using AI for that. Instead, it focuses on the first 72 hours after you discover something is wrong, and on the one-page plan you should write before that day comes.

    One important caveat up front: breach notification is a legal question, and the answer depends on what data was involved, where the affected people live, and what laws and contracts apply to your organization. Nothing here is legal advice. The goal is to help you understand the landscape well enough to make good early decisions, to know when to bring in counsel (early, almost always), and to avoid the mistakes that make a manageable incident much worse.

    Incident or Breach? The Four Ways Day One Usually Starts

    It helps to separate two words that people use interchangeably. A security incident is any event that threatens your systems or data: a malware infection, a suspicious login, a lost device. A breach, in the legal sense that triggers notification duties, is narrower. Under most state laws it means unauthorized access to or acquisition of specific categories of personal information. Under HIPAA it means an impermissible use or disclosure of protected health information that compromises its security or privacy. Every breach starts as an incident, but many incidents never become reportable breaches. Your job in the first 72 hours is to find out which one you have, while protecting your ability to answer that question accurately.

    For small nonprofits, day one almost always starts in one of four ways. Each has a different shape, and recognizing which one you are in tells you what to do first.

    Ransomware

    Files encrypted, a ransom note on the screen

    The loudest scenario. Systems stop working and someone demands payment. Modern ransomware groups usually steal data before encrypting it, so even if your backups are perfect, you may still have a breach. The first priority is isolating affected machines from the network without powering them off or wiping them.

    Compromised email account

    A staff mailbox sending strange messages or rerouting payments

    The most common scenario for small organizations, and the most underestimated. An attacker with access to a mailbox can read years of messages and attachments containing donor, client, or employee data, and may already be using it to redirect a grant payment or vendor invoice. Our guide to AI phishing and wire fraud covers the payment side.

    Lost or stolen device

    A laptop left on a train, a phone taken from a car

    Often the quietest scenario, and the one where preparation pays off most clearly. If the device was fully encrypted and the encryption key was not compromised, many state laws and HIPAA treat the data as secured, which may mean no notification is required. If it was not encrypted, you may have a breach even if nobody ever opens the files.

    Vendor breach notice

    A letter or email from a software provider

    Your systems are fine, but your donor database host, payment processor, or case management vendor was breached. Depending on the law and your contracts, the vendor may notify individuals, or you may be the one legally responsible for notifying them. Do not assume the vendor will handle everything.

    Notice what these scenarios have in common: in none of them do you know, on day one, exactly what data was exposed or whose it was. That uncertainty is normal. The mistake is to resolve it by guessing, either by declaring early that "no data was taken" or by sending panicked notices to everyone before you know what happened. Both create problems you will have to walk back later. The rest of this article is about reducing that uncertainty in a disciplined order.

    Hour Zero: The Calls You Make Before You Touch Anything

    The single most valuable thing a small nonprofit can have on day one is a short list of phone numbers it does not have to look up. Panic makes people do things, and the first things people do on their own tend to be wiping infected machines, resetting everything, or reaching out to the attacker. A few phone calls made in the right order, before anyone starts fixing things, protect your evidence, your insurance coverage, and your legal position.

    The first-call order for most small nonprofits

    Adjust to your policy terms and contracts, but decide the order in advance

    • 1. Your cyber insurer's breach hotline. Most cyber policies require prompt notice and many require you to use the insurer's approved panel of breach lawyers and forensic firms. Hiring your own vendors first can leave those costs uncovered. The hotline will typically connect you with a breach coach (usually an attorney) within hours.
    • 2. Outside counsel with breach experience. If you have no insurance, or your insurer's counsel is not yet engaged, call a lawyer who handles data breaches. Counsel can retain the forensic investigator so that the investigation is conducted at counsel's direction, which may help protect privilege over sensitive findings. Your general nonprofit attorney may not be the right person, but they can refer you.
    • 3. Your IT provider or managed service provider. They know your systems and can isolate devices, disable accounts, and pull logs. Tell them clearly that the priority is containment with evidence preserved, not a quick rebuild.
    • 4. Your bank, if money moved. In email compromise cases involving a fraudulent wire or ACH transfer, call the bank's fraud line immediately and ask about a recall. Speed matters enormously for recovering funds.
    • 5. Your executive director and board chair. Leadership needs to know early, even if details are thin. Decide in advance who has authority to approve spending on response vendors after hours.

    The order between the insurer and counsel matters less than making both calls before the forensic work begins. Many small nonprofits discover only during an incident that their policy has a list of approved vendors and a notice requirement measured in days. Reading your policy's incident response section now, and copying the hotline number into your plan, takes fifteen minutes and can be worth the full value of the coverage.

    Law enforcement is also part of the picture, though it is rarely the very first call. The FBI encourages victims of ransomware and business email compromise to report to the Internet Crime Complaint Center (IC3) and to contact their local FBI field office, particularly when money has been transferred, since early reports can help with fund recovery efforts. CISA also accepts voluntary incident reports through its reporting portal. Reporting does not obligate you to anything, and the information you share can help warn other organizations. Coordinate the timing and content with counsel, but do not let the report slip off the list.

    If ransomware is involved, one more point belongs in your plan: do not contact the attackers or agree to pay anything on your own. Ransom payment decisions involve insurance, sanctions screening under US Treasury rules, and negotiation expertise that your breach coach can arrange. Many organizations that recover from ransomware never pay at all, because their backups were intact and their counsel advised against it.

    Contain the Damage Without Destroying the Evidence

    Containment and evidence preservation pull in opposite directions, and that tension is the heart of the first 24 hours. You want to stop the attacker immediately. You also need to know later exactly what they accessed, because that answer determines who you must notify. Logs, memory, and file timestamps are how forensic investigators answer that question. A machine that has been wiped and reinstalled cannot tell anyone anything, and an organization that cannot show what was accessed may have to assume the worst and notify everyone whose data was on the system.

    The FTC's Data Breach Response: A Guide for Business organizes the early work into securing your operations, fixing vulnerabilities, and notifying appropriate parties, and it stresses taking affected equipment offline while avoiding destroying evidence. NIST's updated incident response guidance, SP 800-61 Revision 3, published in April 2025, frames response as part of an organization's broader cybersecurity risk management rather than an isolated event. For a small nonprofit, the practical translation is a short list of do's and don'ts.

    Do

    Contain while keeping the record intact

    • Disconnect affected devices from the network (unplug the cable, turn off Wi-Fi) but leave them powered on
    • Reset passwords and revoke active sessions for compromised accounts, and check for attacker-created mail forwarding rules
    • Export and save sign-in and audit logs from Microsoft 365 or Google Workspace before retention windows expire
    • Photograph ransom notes and error screens, and keep a written timeline of who did what and when
    • Confirm your backups are isolated and untouched before anyone restores from them

    Don't

    Common moves that make the incident worse

    • Wipe, reimage, or "clean up" machines before forensics has what it needs
    • Delete suspicious emails, rules, or files, which destroys evidence of what the attacker did
    • Discuss the incident over the email system you believe is compromised
    • Speculate in writing about fault or what data was "definitely" taken
    • Post on social media or answer reporters before counsel has reviewed a statement

    Two of these deserve extra attention for small nonprofits. First, log retention: many cloud email and productivity platforms keep detailed audit logs for a limited period, and the length depends on your license tier. If an attacker was in a mailbox for months, the evidence of what they read may be aging out while you decide what to do. Exporting logs is often the single most valuable thing your IT provider can do on day one. Second, out-of-band communication: if email is compromised, set up a temporary channel (a group text, a phone bridge, or personal accounts) so your response team is not discussing the investigation where the attacker can read it.

    Keep the timeline from the very first minute. A simple shared document noting each action, who took it, and when becomes invaluable later. It helps forensics reconstruct events, gives counsel what they need to determine your notification deadlines (which often run from the date of discovery), and shows regulators and insurers that you acted reasonably.

    Figuring Out What Data Was Involved, and Whose

    Every notification decision turns on two questions: what kinds of data were exposed, and where do the affected people live. State breach laws do not apply to all personal data. They typically apply to a person's name combined with specific sensitive elements, such as a Social Security number, driver's license or state ID number, or a financial account or card number together with the code needed to access it. Many states have expanded their definitions to include medical information, health insurance information, biometric data, or online account credentials. A spreadsheet of donor names and mailing addresses alone may not trigger notice under many state laws. The same spreadsheet with bank account numbers from recurring gift forms very likely does.

    This is where a data inventory, prepared in advance, transforms your response. If you already know that your donor CRM holds names, addresses, gift histories, and the last four digits of cards; that your program team's shared drive holds client intake forms with dates of birth and Social Security numbers; and that your HR folder holds W-4s and direct deposit information, then you can quickly tell forensics and counsel where the sensitive data lives and which systems matter most. Without an inventory, someone has to discover all of that during the crisis, often by opening thousands of files by hand.

    Where it lives

    List each system (CRM, email, shared drives, accounting, case management, payroll) and the categories of personal data in each. Note which vendors host it.

    Whose it is

    Donors, clients, employees, volunteers, board members. Note which groups include minors, health information, or people in many states, since each changes the analysis.

    How it is protected

    Is it encrypted at rest? On laptops? In backups? Encryption can be the difference between a reportable breach and an incident that requires no notice.

    For a compromised mailbox, the data question is especially hard. Forensic firms often perform what is called data mining or document review: they collect every message and attachment the attacker could have accessed, then search them for sensitive elements and build a list of affected individuals with their states of residence. This is frequently the slowest and most expensive part of a small nonprofit's breach response, and it is why it matters which mailbox was compromised. A development director's inbox full of donor correspondence and pledge forms will produce a very different list than a communications coordinator's inbox full of newsletter drafts.

    Location matters because state breach laws generally protect residents of the state, not organizations based there. A Chicago nonprofit whose online donors live in 38 states may have obligations under 38 different laws. Building a sense of your geographic footprint in advance, even roughly, helps counsel estimate the scope of notification early. Our overview of state privacy laws and AI explains how this patchwork of state rules affects nonprofits more broadly.

    The US Notification Map: State Laws, HIPAA, and the FTC

    Once you have a working picture of what was exposed, counsel will map it against the laws and contracts that apply. You do not need to master these rules yourself, but understanding their general shape helps you ask good questions, budget realistically, and avoid missing a deadline because nobody realized the clock had started. There are three main layers for most US nonprofits, plus contractual obligations that are easy to forget.

    State breach notification laws

    Apply to nearly every nonprofit that holds personal information

    According to the National Conference of State Legislatures, all 50 states, the District of Columbia, Guam, Puerto Rico, and the US Virgin Islands have laws requiring notification of security breaches involving personal information. They share a basic structure but differ in important details.

    • Timing: most require notice "without unreasonable delay," and a number of states set specific outer limits, commonly 30, 45, or 60 days
    • Regulator notice: many states require notice to the state attorney general or another agency, often when a breach affects more than a set number of residents
    • Content: some states specify what the letter must say, such as a description of the incident, the types of data involved, and contact information for credit bureaus and the FTC
    • Exceptions: most exempt encrypted data where the key was not compromised, and some allow a risk-of-harm analysis that can avoid notice
    • Credit bureaus: large breaches often trigger notice to the nationwide consumer reporting agencies

    HIPAA Breach Notification Rule

    For covered entities and their business associates

    Community health centers, clinics, some behavioral health and substance use programs, and nonprofits that handle protected health information on behalf of covered entities (as business associates) fall under HIPAA. Under the HHS Breach Notification Rule, an impermissible use or disclosure of unsecured PHI is presumed to be a breach unless a documented risk assessment shows a low probability that the information was compromised.

    • Individuals must be notified without unreasonable delay and no later than 60 calendar days after discovery
    • Breaches affecting 500 or more individuals must be reported to HHS within the same 60-day window, and prominent media outlets must be notified when more than 500 residents of a state or jurisdiction are affected
    • Smaller breaches are logged and reported to HHS within 60 days after the end of the calendar year
    • Business associates must notify the covered entity, which generally handles notice to individuals

    HHS has also stated in its ransomware fact sheet that encryption of PHI by ransomware is presumed to be a breach unless the entity can demonstrate a low probability of compromise. For more on HIPAA obligations in an AI context, see our guide to healthcare data protection and AI.

    FTC Health Breach Notification Rule

    For health apps and personal health records outside HIPAA

    Most nonprofits will not fall under this rule, but some will. The FTC Health Breach Notification Rule, amended in 2024, covers vendors of personal health records and related entities that are not covered by HIPAA, which explicitly includes many health and wellness apps. If your organization offers a symptom tracker, wellness app, or similar tool that collects health information from multiple sources directly from consumers, ask counsel whether it applies. The rule requires notice to individuals within 60 days of discovery, notice to the FTC (at the same time for breaches of 500 or more people), and in some cases media notice. Notably, the FTC treats unauthorized disclosures, such as sharing health data with advertising platforms without authorization, as breaches too.

    Contracts, grants, and other obligations

    The layer small nonprofits most often miss

    Government grants and contracts, data sharing agreements with school districts or health systems, and agreements with payment processors often contain their own incident reporting clauses, some with deadlines measured in hours or a few days rather than weeks. Your cyber insurance policy has a notice requirement of its own. Pull these agreements into a single list during planning, note each reporting deadline and contact, and include the list in your plan. Student data has its own considerations; our article on student privacy and FERPA covers that context.

    The practical lesson is that the decision of whether, when, and how to notify belongs to counsel, informed by facts that your team and forensics provide. Your role is to get those facts quickly and accurately, keep the timeline, and make sure nobody has made public commitments ("no personal data was affected") that the investigation later contradicts. Deadlines often run from the date of discovery, which may be the day an employee noticed something odd, not the day leadership was told. That is one more reason to have staff escalate suspicious activity immediately.

    Telling Donors, Clients, Staff, and the Board

    Legal notification letters are one piece of communication. The people who trust your organization deserve more than a compliance document, and how you communicate often determines whether a breach becomes a lasting reputational problem. Nonprofits have an advantage here that companies often lack: an existing relationship built on mission and trust. A breach handled with candor and care can actually reinforce that relationship. A breach handled with silence, spin, or confusion damages it.

    Your board should hear early, in a short factual briefing: what happened, what you know and do not know yet, who has been engaged, and when you will update them. Boards have oversight duties, may need to approve unbudgeted spending, and should never learn about a breach from the news. Our guide to communicating AI risks to your board offers a useful frame for these conversations. Staff need clear instructions about what to say if asked, where to route media or donor calls, and what to do differently while systems are restored. Rumors inside the organization tend to leak outside it.

    Donors and clients who receive formal notice will have questions, and the best response is a prepared FAQ and a designated contact person. Plain language matters. Explain what happened, what information was involved, what you have done about it, what you are offering (such as credit monitoring when financial or identity data was exposed), and what they can do to protect themselves. For clients in vulnerable situations, such as survivors of domestic violence, immigrants, or people receiving sensitive services, consider whether a phone call from a trusted staff member should accompany or precede a letter, and ask counsel how to do that within the law.

    Principles for breach communications

    • Say only what you know. "We are investigating" is honest; "no data was taken" on day two often is not.
    • Have counsel review every external statement, including social posts and donor emails.
    • Use one spokesperson, and route all media and donor inquiries to them.
    • Tell people what they can do, not only what happened.
    • Follow up when the investigation concludes, even if the news is that fewer people were affected than feared.

    Our broader guide to crisis communications with AI covers holding statements, media handling, and message discipline in more depth. The key difference with a breach is the legal overlay: every statement may later be read by a regulator or a plaintiff's lawyer, so speed must be balanced with review.

    Where AI Helps, and Where It Must Stay Out

    AI tools can make a small nonprofit dramatically better prepared for a breach, mostly through work done long before the bad day arrives. They are much less useful, and in some ways dangerous, in the middle of an active incident. The dividing line is simple: AI is excellent for drafting, organizing, and summarizing material that is not itself sensitive, and for building plans and templates that humans and lawyers will review. It should never receive breached personal data in a public tool, and it should never be the one deciding your legal obligations.

    Good uses

    Mostly preparation, always with human review

    • Drafting your one-page incident response plan and role descriptions from a short interview about your organization
    • Generating realistic tabletop exercise scenarios and "injects" tailored to your systems and programs
    • Building a data inventory by turning a walkthrough of your systems into a structured table
    • Drafting template notification letters, donor FAQs, and staff talking points for counsel to review and adapt
    • Helping your IT provider summarize or explain log exports in an approved, enterprise tool, with personal data removed where possible
    • Turning your incident timeline into a clear board briefing

    Never

    Uses that create a second incident or legal risk

    • Pasting breached records, client files, or donor lists into a public or consumer AI tool to "see what's in there"
    • Asking a chatbot whether you are legally required to notify, and treating its answer as the decision
    • Relying on AI-generated summaries of state law deadlines without counsel verifying them
    • Sending AI-drafted notices or public statements without legal review
    • Using AI tools connected to compromised accounts or systems during the response

    The reason for the hard line on pasting breached data is not abstract. Feeding exposed personal information into a tool your organization has not vetted is itself an unauthorized disclosure, and it can expand the scope of the incident you are trying to contain. It may also waive protections your counsel is trying to preserve. If AI analysis of sensitive material is genuinely needed, let your forensic firm or counsel decide which tools are appropriate; they often have their own approved platforms for document review.

    The reason AI must not decide notification obligations is equally concrete. State laws change regularly, their definitions of personal information differ in small ways that matter, and the analysis depends on facts that only forensics can establish. A language model can produce a confident, well-formatted, outdated, or simply wrong summary of a state's deadline. A missed deadline cannot be fixed later. Use AI to prepare your questions for counsel, not to answer them.

    One good preparation exercise: ask an AI assistant to interview you about your organization's systems, vendors, data types, and staff roles, then produce a draft data inventory and a draft plan. The result will be imperfect, but correcting a draft is much faster than starting from a blank page, and the questions it asks will surface gaps you had not considered, such as whether the program team's laptops are actually encrypted.

    When AI Is the Incident

    A growing share of incidents now involve AI directly, and most response plans written before 2024 do not mention them. IBM's 2025 report found that one in five organizations it studied reported a breach involving shadow AI, meaning AI tools used by staff without organizational approval or oversight. For small nonprofits, two AI-related scenarios are especially likely, and your plan should name both.

    A staff member pasted sensitive data into a chatbot

    The well-meaning shortcut

    A program manager uploads a spreadsheet of client names and case details to a free AI tool to draft a grant report. A development associate pastes a donor list with giving history into a consumer chatbot to segment an appeal. These are not malicious acts, but depending on the data and the tool's terms, they may be unauthorized disclosures. Whether they trigger notice is a question for counsel, and the answer depends on what data was shared, the tool's data handling and retention terms, the type of account used, and whether the data can be deleted.

    Treat these like any other incident: document what was shared, when, through which account and tool, and capture the tool's terms at the time. Check whether chat history and any training opt-outs were on, and delete the conversation where possible, but only after counsel confirms you have captured what they need. Then respond without blame. Staff who fear punishment stop reporting, and unreported shadow AI incidents are worse than reported ones. Our guides to shadow AI in nonprofits and writing an AI acceptable use policy cover prevention.

    Your AI vendor was breached

    A vendor notice with an AI twist

    AI vendors are vendors. If a meeting transcription service, AI writing assistant, or donor analytics platform you use discloses a breach, handle it like any vendor breach notice, with a few extra questions. What did you send the tool? Transcripts of client meetings, board discussions, and HR conversations can contain far more sensitive information than people remember. Did the vendor retain prompts, files, or outputs, and for how long? Were any integrations connected to your email, drive, or CRM, and do those tokens need to be revoked?

    Revoke connected-app permissions immediately, rotate any API keys, and ask the vendor in writing exactly which customer data was affected. Your contract should already say how quickly the vendor must notify you and what cooperation they owe; if it does not, that is a gap to close next time. Our guide to AI vendor security evaluation lists the contract and security questions to ask before signing.

    Writing the One-Page Plan Before You Need It

    A small nonprofit does not need a 60-page incident response manual. It needs a single page that a stressed person can follow at 7 a.m. on a Monday, printed and stored somewhere other than the systems that might be compromised. Large plans tend to go unread. A one-page plan with names and phone numbers tends to get used. You can add appendices later, such as your data inventory and contract reporting list, but the first page should stand on its own.

    What belongs on the page

    Keep a printed copy and an offline digital copy

    • How to report: one sentence telling every staff member and volunteer who to call, immediately, if something seems wrong, and that they will not be blamed for reporting.
    • Roles: an incident lead (often the executive director or operations director), a technical lead (your IT provider), a communications lead, and a board liaison, each with a named backup.
    • Contact list: insurer breach hotline and policy number, outside counsel, IT provider after-hours line, bank fraud line, FBI field office, and key vendor security contacts.
    • First-hour steps: isolate, do not wipe, start the timeline, switch to out-of-band communication, make the calls in order.
    • Authority: who can approve emergency spending and engage vendors, and up to what amount, without waiting for a board meeting.
    • Pointers: where to find the data inventory, the contract reporting list, backups documentation, and template communications.
    • AI rules: no breached data in unapproved AI tools, and no notification decisions without counsel.

    A plan nobody has practiced is a hypothesis. A tabletop exercise is a structured discussion, usually 60 to 90 minutes, where your response team walks through a realistic scenario and talks through what they would do at each step. No systems are touched. The value comes from discovering, in a calm conference room, that the IT provider's after-hours number goes to voicemail, that nobody knows where the insurance policy is, or that the board chair assumed the executive director could approve a $40,000 forensics engagement without a vote.

    You do not need to write scenarios from scratch. CISA publishes free Tabletop Exercise Packages covering ransomware, phishing, insider threats, and other scenarios, designed for organizations to run on their own. AI is genuinely useful here too: give an assistant a description of your organization (with no sensitive data) and ask it to adapt a CISA scenario to your systems, add timed injects such as "a reporter calls asking about the incident" or "the vendor says the data was not encrypted," and draft discussion questions for each stage. Run one exercise a year at minimum, rotate the scenario among the four types described above, and update the one-page plan with whatever you learn.

    A realistic preparation timeline

    About a month of part-time effort for a small team

    • Week 1: Read your cyber policy's incident section, save the hotline number, and identify outside counsel. Ask your IT provider about log retention and backup isolation.
    • Week 2: Build a first-pass data inventory with AI assistance, and list contracts and grants with incident reporting clauses.
    • Week 3: Draft the one-page plan and template communications, then have counsel review the templates.
    • Week 4: Run your first tabletop exercise, brief the board, and print the plan.

    Conclusion

    A data breach is one of the few organizational crises where the first few hours can determine the outcome of the next several months. The nonprofits that come through well are rarely the ones with the biggest security budgets. They are the ones that knew who to call, preserved the evidence that let investigators narrow the scope, understood that notification is a legal decision made on facts, and communicated with donors and clients honestly. None of that requires technical sophistication. It requires a plan and a little practice.

    The legal landscape is complex but navigable. All 50 states have breach notification laws with differing timelines and requirements, HIPAA adds its own rules for covered entities and business associates, and the FTC's Health Breach Notification Rule reaches some health-related services outside HIPAA. Your job is not to master every rule. It is to engage counsel early, give them accurate facts fast, and keep track of every contractual deadline that sits alongside the statutes.

    AI fits into this picture mostly before the incident, as a fast and capable assistant for drafting plans, building inventories, designing tabletop scenarios, and preparing templates for counsel. During an incident, it must stay away from breached data and away from legal judgments. And because AI tools are now part of how staff work, your plan should treat a mistaken chatbot upload or an AI vendor breach as a real incident with a real response. Write the one-page plan this month. The best time to decide who to call first is a quiet afternoon, not the morning your files are locked.

    Be Ready Before Day One

    We help small nonprofits build data inventories, one-page incident response plans, AI use rules, and tabletop exercises that prepare staff and boards for the day something goes wrong.