AI AND NONPROFIT FUNDRAISING IN 2027

The good the bad and the ugly for nonprofit leaders and boards

By Brian Phelps | Transformation Growth | October 2026

As your nonprofit prepares for 2027, AI deserves a place in the conversation about fundraising, staffing, and mission delivery. It can help a small team communicate more consistently and organize work that otherwise gets postponed. It can also multiply mistakes and create serious harm when leaders fail to verify information or protect the people behind their data.

I encourage nonprofit leaders to judge AI by whether it strengthens donor relationships and your ability to serve. Faster appeals are useful only if they are accurate, respectful, and effective. This article uses published findings to establish the current landscape, then offers my outlook and recommendations for 2027. Examples are illustrative; they are not documented incidents or promised results.

The fundraising landscape heading into 2027

Giving USA 2026 estimates that U.S. charitable giving reached $617.20 billion in 2025, up 5.7 percent in current dollars and 3.0 percent after inflation. Individuals contributed $394.2 billion, foundations $117.15 billion, bequests $62.19 billion, and corporations $43.67 billion. These are national estimates for 2025, not a forecast of your organization’s 2027 revenue. [1]

A separate view comes from the Fundraising Effectiveness Project’s first quarter 2026 report. It estimated a 4.3 percent year over year increase in dollars raised alongside a 0.8 percent decline in donor count. The dataset covered approximately 15,700 organizations, 3.2 million donors, and $3.5 billion. The report uses an updated methodology and does not capture every form of giving or local influence. Its quarterly results should not be compared directly with a full year national total. [2]

My interpretation is that a higher fundraising total can conceal a weakening donor base. An organization can meet its budget through several large gifts while losing many smaller supporters. For 2027, leaders should watch the breadth and durability of support alongside total dollars. AI may help identify gaps in follow up, but someone must still build the relationship.

AI adoption is moving faster than readiness

NTEN and The Bridgespan Group surveyed 917 nonprofit staff and executives in summer 2026. Among 404 executive respondents to capacity questions, about 18 percent reported effective use training already in place, and about 24 percent reported data sufficiently organized to support AI. In governance responses, about 40 percent reported rules on what data could be entered into AI tools. These self reported findings describe participating respondents, not a census of nonprofits. [3]

A fundraising software vendor, Virtuous, surveyed 346 nonprofit organizations in December 2025. Its 2026 report said 92 percent of respondents used AI in some capacity, while only 7 percent reported major improvements in achieving their mission. This vendor survey should not be interpreted as a representative adoption rate across the entire sector or evidence that AI caused mission improvement. [4]

For a small nonprofit, the practical question is whether the team has the information, time, and accountability needed to use a tool well. A modest pilot with clean records may be more useful than an ambitious system operating on incomplete donor histories.

The good - more capacity for meaningful fundraising

AI can help staff get from a blank page to a workable draft. An approved program description can become a first draft of an appeal, a board briefing, or a donor update. Staff can then verify the facts, adjust the tone, and make the request specific. The benefit is the capacity recovered for work that needs personal attention.

Imagine a community food pantry preparing a year end campaign. Staff provide verified information about services, approved costs, and the campaign purpose. AI produces several draft messages for review. The development director checks every figure and explains what the gift will support without implying a precise outcome the organization cannot substantiate.

AI can also support stewardship, the work of acknowledging donors and showing how funds are used. With approved systems and appropriate permissions, it can help flag overdue acknowledgments or prepare drafts for staff review. A thoughtful update about a completed project can give supporters a reason to stay involved between requests.

For grants, AI can help organize a proposal outline, compare a draft with published application requirements, and identify unanswered questions. A staff member must verify eligibility, deadlines, budget calculations, and the funder’s rules about AI use directly from the current application materials. A fluent draft cannot establish that your organization qualifies or that the promised program is feasible.

AI can help interpret donor patterns too, but begin with reliable definitions and records. Separate first time donors, repeat donors, recurring supporters, and lapsed donors. Use observed giving and consented interests to plan relevant outreach. Treat suggested priorities as hypotheses to review rather than predictions of what a person will give.

The bad - everyday mistakes that weaken results

The most common operational problem to guard against is plausible but inaccurate output. A draft may invent a statistic, misstate an application requirement, confuse two donors, or exaggerate the impact of a gift. Every public claim should trace to an approved source: a financial record, program report, current funder instruction, or consented story.

Automation can also make communications feel impersonal. A message that inserts someone’s name is not necessarily relevant to their interests. If a longtime supporter receives a generic welcome appeal, the organization signals that it has not paid attention. Review both the message and the audience selection before sending.

Poor data can distort outreach. Duplicate records may cause repeated requests; incomplete gift histories may classify a loyal donor as lapsed. A model trained on past fundraising decisions may keep prioritizing people already well represented in your database. My recommendation is to review exclusions as carefully as the top ranked prospects and preserve opportunities for new supporters to participate.

Costs deserve attention as well. Include subscription fees, configuration, training, review, and correction time. If staff spend as long checking a draft as they once spent writing it, the process may need to change. A saved hour is not a financial saving unless the organization actually reduces an expense or uses the capacity productively.

Leadership can compound these problems by treating AI as a reason to increase volume without examining workload. More drafts can create a larger approval queue. More appeals can create donor fatigue. Define a useful outcome before increasing output, and involve the people responsible for the resulting work.

The ugly - when AI damages dignity and public trust

Fabricated beneficiary stories are a serious boundary. Do not invent a family, quotation, outcome, or testimonial and present it as real. A generated image should not be represented as a photograph of someone your organization served. If you use an illustration or clearly labeled hypothetical example, make its nature unmistakable. Use actual stories only with appropriate permission and attention to the person’s dignity.

Consider an AI generated appeal describing a child who received help from your organization. If the child and circumstances are invented but the appeal presents them as factual, the problem goes beyond poor writing. The nonprofit is asking for support using a false representation. A compelling emotional response does not justify it.

Privacy failures can be equally damaging. Donor records may contain addresses, financial information, family circumstances, and confidential notes. Beneficiary records may be even more sensitive. Before any upload, confirm that the tool and proposed use are approved, review access and retention controls, and limit the information to what the task requires. Removing a name alone may not prevent identification.

The Association of Fundraising Professionals’ ethical standards emphasize truthful communications, protection of confidential information, respect for donor intent, and responsible reporting on funds. The code governs AFP members and provides a useful professional reference; it is not a substitute for applicable law or your organization’s policies. AI use should meet those commitments. [5]

Fraud is another real concern. In a January 2025 alert, the FBI warned that AI could make charitable fraud or impersonation appear more legitimate. It reported more than 4,500 complaints and approximately $96 million in 2024 losses involving fraudulent charities, crowdfunding accounts, and disaster relief campaigns. Those figures cover the broader fraud categories, not losses proven to be caused by AI. [6]

For your organization, publish a clear official donation route and help donors confirm unusual requests. Require independent verification of changes to payment instructions and urgent transfer requests, using a known contact method. A familiar sounding voice or realistic video should not by itself authorize a financial action.

If a serious error or suspected exposure occurs, pause the affected process, preserve records, and involve the responsible leaders and advisers. Correct false public claims promptly. Assess who was affected and what notification or other response is required. An organization that quietly leaves a false story in circulation compounds the harm.

Keep the relationship at the center

I often encourage nonprofits to ask whether an activity helps attract a new donor or keep the donors they have. AI can support that work by reducing administrative friction. It should also leave room for listening: Why did this person give? What do they want to understand? How do they prefer to hear from us?

A useful practice is to assign recovered time to specific relationship work. If drafting an update becomes easier, use part of the available time to call supporters, thank volunteers, or ask a first time donor about their interests. Decide who will do that work and when. Otherwise the time may simply disappear into another batch of content.

My 2027 outlook is that credible reporting will become more valuable as polished fundraising material becomes easier to produce. Explain what you accomplished, what remains unfinished, how contributions were used, and what comes next. Distinguish people served from services delivered, and separate funds received from pledges. Help donors understand the reality behind the appeal.

A practical plan for 2027

·         Choose one manageable use such as drafting a donor update from approved public information. Establish a baseline for time and quality before starting.

·         Name the accountable person and approver. Clarify what AI may draft, what it may access, and which actions require human authorization.

·         Create written rules for data, claims, stories, images, and donor communications. Include volunteers and contractors who perform the work.

·         Build a verified source set from program reports, approved budgets, mission language, and current campaign facts. Resolve inconsistent numbers before drafting.

·         Test before expanding. Review actual errors, staff workload, donor feedback, and the full cost. Correct problems before automating additional steps.

·         Report useful results to the board. Measure fundraising performance and mission consequences, not simply how many people used AI.

A fundraising dashboard can track donor retention, first to second gift conversion, recurring donor retention, reactivated donors, net campaign proceeds after direct costs, and acknowledgment turnaround time. Keep definitions consistent. For annual donor retention, divide prior year donors who also gave this year by all prior year donors. Compare like periods and distinguish campaign cash receipts from pledges. Do not attribute every change to AI without a credible comparison.

The leadership choice

The "good" is additional capacity to communicate clearly and follow through. The "bad" is faster production of mistakes, irrelevant messages, and work that costs more than it delivers. The "ugly" is deception, exposure of private information, exploitation of vulnerable people, and fraud that destroys confidence in the mission.

For 2027, I encourage boards and nonprofit leaders to make deliberate choices about each use. Put accountability beside opportunity. Give your team permission to learn and the guidance to recognize boundaries. The strongest result is a donor who understands your work, trusts your organization, and chooses to remain part of it.

Sources and research notes

[1] Indiana University Lilly Family School of Philanthropy. Giving USA 2026 release June 23 2026. National estimates for 2025. https://philanthropy.indianapolis.iu.edu/news-events/news/_news/2026/giving-usa-report-2026.html

[2] Fundraising Effectiveness Project. 2026 Q1 report and dataset overview. https://publications.fepreports.org/

[3] NTEN and The Bridgespan Group. State of Nonprofit AI Adoption and Governance September 10 2026. Executive response figures are rounded. https://www.nten.org/publications/state-of-nonprofit-ai

[4] Virtuous. 2026 Nonprofit AI Adoption Report February 16 2026. Vendor survey conducted December 2025. https://virtuous.org/blog/2026-nonprofit-ai-adoption-report/

[5] Association of Fundraising Professionals. Code of Ethical Standards updated December 2023. https://afpglobal.org/ethics/code-ethical-standards

[6] FBI. Beware of Charitable Fraud Related to Mass Casualty and Disaster Events January 16 2025. https://www.fbi.gov/investigate/cyber/alerts/2025/beware-of-charitable-fraud-related-to-mass-casualty-and-disaster-events

Research reviewed October 10 2026. The 2027 outlook, scenarios, and implementation suggestions are the author’s interpretation. No cited survey establishes that AI will increase an individual nonprofit’s fundraising revenue.