AI Social Engineering Fraud for Nonprofits, Municipalities, and Community Organizations
A nonprofit executive director receives an urgent email from the board chair requesting a grant payment. A municipal employee receives a phone call that sounds exactly like the finance director. An HOA treasurer receives a vendor invoice with new banking instructions. A church administrator gets a message from a familiar donor asking for a confidential refund.
Each request looks plausible. Each one may be powered by artificial intelligence.
AI has changed social engineering from a recognizable email problem into a multi-channel impersonation threat. Criminals can create polished emails, clone voices, imitate writing styles, produce convincing invoices, register look-alike domains, and use information from public websites to target small organizations with remarkable precision.
For mission-driven organizations, the concern is not only whether someone can hack a network. The concern is whether a trusted person can be manipulated into authorizing a payment, changing payroll information, releasing donor data, or bypassing a normal approval process.
This article is a spoke in our broader series on AI-enabled fraud and insurance. Read the pillar article, AI-Powered Social Engineering Fraud: Is Your Business Insurance Keeping Pace in 2026? for the wider risk and insurance framework.
Why “we’re a small nonprofit” is no longer a safe assumption
Small organizations often believe criminals target large corporations, banks, or government agencies. That assumption is no longer reliable.
Attackers target organizations that:
Receive grants, donations, tuition, dues, or public funds
Pay vendors and contractors by ACH or wire
Use part-time finance personnel or volunteers
Depend on a small number of people with broad authority
Operate with informal approval processes
Publicly list executives, board members, employees, vendors, and donors
Have limited cybersecurity staff or technology budgets
Need to move quickly during emergencies or community programs
A small nonprofit may have fewer employees, but it can still control significant grant funds, payroll accounts, donor information, and program money. A local municipality may have formal procedures, but it also processes large vendor payments through departments with different levels of training and oversight.
The FBI’s Internet Crime Complaint Center reported 24,768 business email compromise complaints and more than $3 billion in reported losses in 2025. IC3 also reported more than 22,000 complaints involving an AI component, with approximately $893 million in adjusted losses across reported cybercrime categories. These figures do not separately identify all nonprofit or municipal losses, and they do not capture every incident. They show the scale of the broader threat.
Review the FBI’s Business Email Compromise guidance and the IC3 Business Email Compromise resource for official prevention and reporting information.
How AI-enabled fraud affects mission-driven organizations
Traditional phishing often contained obvious warning signs: poor grammar, unfamiliar senders, strange formatting, and generic requests. AI makes those clues less dependable.
A fraudulent request may now:
Use the correct name of the executive director
Refer to a real grant, project, event, or vendor
Match the sender’s writing style
Include a realistic-looking invoice or letterhead
Arrive through email, text message, phone, or a video meeting
Use a cloned voice to create urgency
Follow an earlier conversation stolen from a compromised mailbox
Ask the recipient to keep the request confidential
Use a look-alike domain that differs by only one character
The attack usually exploits trust rather than a technical vulnerability. The criminal wants the recipient to believe the request came from someone who has authority.
A simple definition: social engineering
Social engineering is the use of deception or psychological pressure to persuade a person to disclose information, approve an action, or transfer money.
A simple definition: business email compromise
Business email compromise, or BEC, is a fraud scheme involving a compromised or impersonated email account to induce an unauthorized transfer of money or other valuable assets.
AI expands both threats because the criminal can make the impersonation more believable and conduct more conversations at scale.
The most common fraud scenarios
1. Grant fraud and grant-related payment scams
Nonprofits depend on grants from government agencies, foundations, corporations, and private donors. Attackers use that dependence to create urgency.
A fraudster may impersonate:
A foundation program officer
A federal or state grant administrator
A nonprofit’s grant consultant
A partner organization
A major donor
A government agency requesting a compliance payment
The organization may receive a fake award notice, a request to pay a processing fee, or a demand to send funds to a “temporary” account. The message may include actual program names and grant terminology copied from public documents.
Grant fraud can create several types of loss:
Money sent to a fraudulent account
Fees paid to obtain a nonexistent grant
Donor or funder funds redirected to a criminal
Sensitive grant application information disclosed
Costs incurred while investigating the fraud
Regulatory, contractual, or reputational consequences
Insurance does not automatically cover the value of an expected grant that never existed. It may also not cover every payment made in response to a fraudulent grant communication. Coverage depends on the policy language, the definition of a covered loss, the method of transfer, and any social engineering or fraudulent instruction endorsement.
2. Vendor payment diversion
Vendor payment diversion is one of the most practical risks for nonprofits, municipalities, HOAs, healthcare organizations, and religious institutions.
The attacker impersonates a legitimate vendor and asks the organization to update its banking information. The request may arrive after the attacker compromises either the vendor’s email account or the organization’s account payable mailbox.
Common targets include:
Construction contractors
Cleaning and maintenance companies
Information technology providers
Medical suppliers
Utility providers
Property management companies
Consultants
Transportation providers
Grant partners and subcontractors
The organization updates its records and sends the next payment to the criminal. The legitimate vendor later reports that it was never paid.
A changed bank account should never be approved solely because the request came from a familiar email address. Verify the request using contact information already stored in your records. Do not use the telephone number contained in the new email or invoice.

3. Fake donor emails and donation diversion
A fraudster may impersonate a donor who wants to make a large contribution, request a refund, or change the destination of a charitable payment.
Examples include:
A fake donor sends a counterfeit check and asks the nonprofit to return part of the money
A criminal impersonates a donor and requests a refund to a different account
A real donor’s email account is compromised and the attacker changes payment instructions
A fraudulent donation platform captures donor payment information
A fake corporate sponsor requests sensitive tax or banking documents
Organizations should be cautious when a new donor wants an unusually large transaction, demands rapid processing, or requests a refund before a payment fully clears.
4. Payroll direct-deposit fraud
Payroll fraud does not always require access to the payroll system. An attacker may impersonate an employee and ask human resources or a payroll administrator to change direct-deposit information.
The request may look like:
“I changed banks and need this updated before Friday.”
“I am traveling and cannot complete the form.”
“Please handle this confidentially.”
“The employee portal is not working, so I am sending the new account details here.”
A single change may affect one employee. A compromised payroll administrator’s account may allow the attacker to redirect multiple employees’ paychecks.
Use a documented process that requires:
Employee verification through a known telephone number
A secure employee portal or signed form
Review by a second authorized person
Confirmation after the change is entered
Temporary monitoring for unusual payroll activity
5. Deepfake executive, mayor, pastor, or board-chair instructions
Voice cloning makes a familiar voice an unreliable authentication method by itself.
A criminal may call while imitating:
A mayor or town manager
A finance director
An executive director
A board chair
A pastor or religious leader
A physician or healthcare administrator
An HOA property manager
A department head
The request usually involves urgency, secrecy, or an unusual payment. It may include a follow-up text message or email that reinforces the instruction.
The correct response is not to embarrass the caller or challenge the person’s identity aggressively. The correct response is to follow the organization’s verification procedure. Call the person back using a known number. Ask another authorized person to confirm the request. Never treat a familiar voice as sufficient approval for a financial transfer.
How organizational structure changes the risk
Mission-driven organizations often have more complex authority structures than they realize.
A nonprofit may have:
A paid executive director
A volunteer treasurer
A board finance committee
A part-time bookkeeper
An outside accountant
A grant administrator
Multiple program managers
Volunteers with access to donor or payment platforms
A municipality may have:
Elected officials
A town or city manager
A finance department
Procurement personnel
Department heads
School or public safety departments
Outside contractors
Shared services with another municipality
An HOA may rely on:
Volunteer board members
A property management company
An outside accountant
A management committee
Several authorized signers
A small number of residents handling sensitive records
These arrangements can create unclear boundaries. One person may add a vendor, another may approve the payment, and a third may release the funds. That separation can help prevent fraud, but only if each person understands the process and independently verifies the transaction.
The presence of multiple signers does not automatically eliminate risk. If two signers rely on the same fraudulent email, both approvals may reinforce the mistake.
Which insurance policies may be relevant?
No single policy automatically addresses every AI-enabled fraud event. A coverage review should examine the entire insurance and bonding program.
Cyber insurance
Cyber liability insurance may address data breaches, network security incidents, privacy claims, business interruption, incident response, and certain forms of cybercrime.
For social engineering, ask whether the policy specifically addresses:
Social engineering fraud
Funds transfer fraud
Fraudulent instruction
Business email compromise
Payment diversion
Payroll diversion
Vendor impersonation
Deepfake-enabled instructions
Costs to investigate the incident
Coverage for compromised email or cloud accounts
The most important issue is often the sublimit. A policy may have a large overall cyber limit but a much smaller limit for social engineering or fraudulent transfer losses. The sublimit may also have a separate retention and strict verification requirements.
Do not assume that a cyber policy covers a voluntary transfer simply because the transfer resulted from a cyber incident. Some policies require unauthorized access to a computer system. Others provide a specific extension for social engineering. The wording controls.
Crime insurance
Commercial crime insurance, sometimes called fidelity coverage, may address employee dishonesty, theft, forgery, funds transfer fraud, computer fraud, and social engineering. The available coverage depends on the form and endorsements.
Ask these questions:
Does the policy cover losses caused by outside criminals?
Does it cover volunteers?
Does it cover temporary or part-time workers?
Does it cover the organization’s own money?
Does it cover donor, grant, or client funds held in trust?
Does it cover fraudulent changes to vendor banking information?
Does it cover payroll direct-deposit diversion?
Does it cover a payment authorized by an employee who was deceived?
Does it require a callback or dual approval procedure?
Does it apply when the first payment was legitimate but later payments were diverted?
Many organizations purchase crime coverage for employee theft but overlook external social engineering. Those are different risks. Review the wording rather than relying on the policy title.
Faithful performance bonds
A faithful performance bond is a bond intended to protect an organization or public entity against certain dishonest or wrongful acts by a person required to perform duties faithfully. Municipal officials, nonprofit employees, treasurers, and other fiduciaries may be subject to bonding requirements or internal bonding decisions.
A faithful performance bond is not automatically interchangeable with crime insurance.
Important questions include:
Who is bonded?
Is the person an employee, officer, elected official, or volunteer?
What acts trigger the bond?
Does the bond require dishonesty?
Does it cover a mistake caused by deception?
Does it cover an external fraudster’s conduct?
Does it cover funds transferred voluntarily by the organization?
What notice and proof requirements apply?
Does another insurance policy respond first?
If an authorized employee is tricked into sending money, the event may not meet a bond’s definition of dishonest conduct. A bond may respond differently when the bonded person steals funds or knowingly violates a duty. The distinction is highly specific to the bond wording and facts.
Directors and officers insurance
Directors and officers insurance generally addresses claims alleging wrongful acts in the management or governance of an organization. It may help with defense costs and covered settlements, subject to the policy terms.
D&O insurance usually does not replace stolen money. It may become relevant if a board, officer, or executive is accused of:
Failing to supervise financial controls
Mismanaging donor or grant funds
Failing to respond appropriately to a known cyber risk
Neglecting required governance duties
Making misleading statements after a fraud event
Failing to protect confidential information
Breaching fiduciary duties
D&O coverage also contains exclusions and conditions. A fraud allegation, personal profit exclusion, prior knowledge issue, or insured-versus-insured provision may affect the analysis.
General liability and other commercial insurance
General liability insurance is designed primarily for third-party bodily injury, property damage, personal injury, and advertising injury claims. It is not a substitute for cyber, crime, or D&O coverage.
A broader commercial insurance review may also identify related exposures involving:
Professional services
Employment practices
Privacy liability
Property and business income
Equipment breakdown
Workers’ compensation
Commercial auto
Umbrella liability
The goal is to identify which policy addresses the direct financial loss, which policy addresses liability claims, and which policy provides incident-response support.
Coverage comparison at a glance
Scenario | Coverage that may be relevant | Key issue to verify |
Vendor bank account changed after an AI-written email | Crime or cyber social engineering endorsement | Does the policy cover voluntary payments and payment diversion? |
Payroll direct deposit sent to a criminal account | Crime funds transfer coverage or cyber fraud coverage | Does coverage apply to payroll changes and employee impersonation? |
Executive director voice cloned to authorize a wire | Crime or cyber fraudulent instruction coverage | Does the policy require a specific authentication procedure? |
Employee secretly steals grant funds | Crime or fidelity coverage | Are grant funds and the employee included in the insured definition? |
Board accused of failing to maintain controls | D&O | Does the policy provide entity and defense coverage for the alleged wrongful act? |
Donor database exposed during a phishing attack | Cyber liability | What are the breach response, notification, and regulatory limits? |
Fake grant never existed and the organization pays a fee | Possibly limited or no direct-loss coverage | Is the payment a covered fraudulent transfer or an uncovered business loss? |
Municipal official fails to follow required duties | Faithful performance bond, crime, or D&O | Does the bond require dishonesty, and which policy responds first? |
This table is a planning tool, not a coverage determination. Every claim depends on the policy language, facts, exclusions, conditions, limits, and applicable law.
Practical controls for nonprofits, municipalities, HOAs, and community organizations
Require independent verification
Verify every request involving:
A new vendor
A bank account change
A payroll change
A large wire or ACH transaction
An urgent grant payment
A donor refund
A confidential or off-cycle payment
Use a trusted phone number or an in-person conversation. Do not use contact details supplied in the suspicious message.
Separate duties
The person who creates a vendor should not be the only person who approves the payment. The person who enters payroll changes should not be the only person who reviews the payroll register.
Even a small nonprofit can establish a basic separation-of-duties procedure involving the treasurer, executive director, board finance committee, or outside accountant.
Set transaction thresholds
Define when additional review is required. For example:
One approval for routine payments below a defined amount
Two approvals for new vendors or changed bank details
Finance committee or board approval for unusually large transfers
Written documentation for emergency payments
The correct threshold depends on the organization’s finances and operating model.
Use phishing-resistant authentication
Enable multifactor authentication on:
Email
Payroll systems
Accounting software
Banking platforms
Donor management systems
Cloud storage
Grant-management portals
Prioritize phishing-resistant authentication methods for administrators and financial users. Also remove access promptly when an employee, contractor, or volunteer leaves.
Protect email domains
Organizations should discuss SPF, DKIM, and DMARC with their email administrator. These technical controls can reduce some forms of domain spoofing, although they do not prevent every impersonation attack.
Also consider:
Blocking automatic forwarding rules
Reviewing mailbox-login alerts
Restricting administrative privileges
Monitoring suspicious inbox rules
Using password managers
Limiting access to donor and payroll information
Train volunteers, not only employees
A volunteer treasurer, board member, or committee chair may have access to financial systems without receiving formal security training. Include those individuals in annual training and payment-control exercises.
Training should cover AI-specific tactics:
Voice cloning
Fake video meetings
Highly personalized emails
Look-alike domains
Urgency and secrecy
Requests to bypass normal procedures
The goal is not to teach people to identify every deepfake. The goal is to make verification a required process.

What should you do after a fraudulent transfer?
Act immediately. Do not wait to determine whether insurance will respond.
Contact the originating bank and request a recall or reversal.
Contact the receiving bank if instructed by law enforcement or your financial institution.
Notify the organization’s leadership, IT provider, accountant, and legal counsel.
Preserve emails, headers, invoices, phone numbers, payment records, and login information.
Contact the insurance broker or carrier using the policy’s claim-reporting instructions.
Avoid deleting compromised accounts or messages before preserving evidence.
File a report with the FBI’s Internet Crime Complaint Center.
Consider contacting local law enforcement and the appropriate state or federal agency.
The FBI emphasizes that rapid reporting can improve the chance of recovery. It does not guarantee recovery.
Organizations can also review the Connecticut State and Local Cybersecurity Grant Program and the Connecticut Nonprofit Security Grant Program for program information. Grant eligibility, allowable expenses, and application periods change. Do not assume a grant will pay insurance premiums or reimburse a fraud loss.
How much does nonprofit fraud insurance cost?
There is no universal price for nonprofit fraud insurance, municipal cyber insurance, HOA fraud coverage, or a faithful performance bond.
Pricing depends on factors such as:
Annual revenue or operating budget
Payroll and employee count
Grant and donation volume
Payment volume and transaction size
Number of authorized users and signers
Use of volunteers
Banking procedures
Multifactor authentication
Prior losses
Data held by the organization
Coverage limits and sublimits
Retentions or deductibles
Geographic footprint
Contractual or bonding requirements
The cheapest policy may provide inadequate protection if the social engineering sublimit is too low or if it excludes the organization’s most common payment process.
Ask for a side-by-side explanation of:
Total policy limit
Social engineering sublimit
Funds transfer fraud limit
Computer fraud limit
Crime retention
Cyber retention
Coverage for volunteers
Coverage for funds held for others
Verification-condition requirements
Claim reporting requirements
Exclusions for voluntary transfers
Coverage for investigation and response expenses
A broker should explain the differences in plain language and identify where coverage remains uncertain.
Questions to ask before buying coverage
Use these questions during a nonprofit, municipal, HOA, healthcare, or community organization insurance review:
If an employee receives a deepfake voice call and authorizes a wire, which policy may respond?
If a legitimate vendor’s email is compromised, is the resulting payment diversion addressed?
Is social engineering coverage part of the main limit or a separate sublimit?
Does coverage apply to ACH, wire, check, and payment-platform transactions?
Are volunteers included in crime and fidelity definitions?
Are donor, grant, patient, resident, or client funds covered?
What controls must the organization follow before a claim is paid?
Does the bond protect against external fraud, internal dishonesty, or both?
How does D&O respond if the board is accused of poor oversight?
What must the organization do immediately after discovering a loss?
For a second industry perspective, you can review nonprofit cybersecurity resources from ICONN Insurance Solutions. You can also read community discussions about business email compromise on Reddit’s cybersecurity community. Reddit posts are user-generated and should not replace advice from your broker, attorney, accountant, bank, or law enforcement.
The FBI’s public educational videos and channel are available through YouTube. Use official sources when reviewing fraud-prevention guidance.
Frequently asked questions
Does cyber insurance cover social engineering fraud?
Sometimes. Cyber policies may include social engineering, fraudulent instruction, or funds transfer coverage, but those extensions often have separate sublimits, retentions, conditions, and exclusions. Review the actual policy wording.
What is nonprofit fraud insurance?
“Nonprofit fraud insurance” usually refers to crime, fidelity, funds transfer, computer fraud, or social engineering coverage designed to address financial losses caused by dishonest or fraudulent acts. The exact coverage depends on the policy form.
Do small nonprofits need crime coverage?
Size does not eliminate exposure. A small nonprofit may control grants, donations, payroll, and sensitive data. Crime coverage should reflect the organization’s actual payment activity and internal controls.
Does a faithful performance bond cover a deepfake scam?
Not automatically. A faithful performance bond may focus on dishonest or wrongful acts by a bonded person. A payment made because an employee was deceived may be treated differently from an intentional theft or knowing breach of duty.
Does D&O insurance reimburse stolen funds?
D&O insurance generally addresses covered management liability claims rather than the direct loss of stolen money. Crime, fidelity, or cyber fraud coverage may be more relevant to the direct financial loss.
Is a municipality protected because it has multiple payment signers?
No. Multiple signers help only when each signer independently follows verification and approval procedures. Two people can approve the same fraudulent request if both rely on the same compromised communication.
What is the first step after a fraudulent wire transfer?
Contact the bank immediately and request a recall or reversal. Then preserve evidence, notify relevant internal and insurance contacts, and report the incident to IC3. Recovery is not guaranteed, so speed matters.
How can an HOA improve fraud controls?
Require two-person approval for payment changes, verify vendor instructions through a known phone number, restrict banking access, train volunteer board members, and review crime or fidelity coverage for HOA-specific exposures. Insure Connecticut’s HOA insurance article provides additional context on financial protection for associations.
Final takeaway
AI-enabled social engineering has made financial fraud more convincing, more personal, and more difficult to detect through visual clues alone. Nonprofits, municipalities, healthcare organizations, HOAs, religious institutions, and community groups face exposure whenever people can authorize payments, change banking instructions, manage payroll, or access donor and client information.
The strongest response combines:
Independent payment verification
Separation of duties
Multifactor authentication
Volunteer and employee training
Clear incident-response procedures
Appropriate cyber and crime coverage
Careful review of D&O and faithful performance bonds
Request a commercial insurance review from Insure Connecticut LLC to evaluate your organization’s cyber, crime, management liability, and bonding exposures across Connecticut, New York, Massachusetts, Rhode Island, New Jersey, and other states where coverage is available. Coverage is subject to policy terms, conditions, exclusions, limits, and carrier underwriting.
.png)

Comments