Phone Fundraising in 2026: Where AI Belongs on a Call Sheet

Why Phone Fundraising Still Earns Its Place

"Phone fundraising" covers several distinct programs with different goals, and AI helps each of them in slightly different ways. A thank-a-thon is a stewardship call with no ask at all: staff, board members, or volunteers call recent donors simply to say thank you. A lapsed donor reactivation campaign calls people who gave in the past but not recently, to reconnect and, gently, to invite a renewed gift. Monthly giving upgrade calls go to existing sustainers or loyal annual donors to invite them to start or increase a recurring gift. Board call nights put trustees on the phone, usually for thank-yous or year-end renewals. Colleges and independent schools run student and alumni telefunds, where paid or volunteer student callers reach alumni and parents. And many organizations hire professional telefundraising firms to run large calling campaigns on their behalf.

The evidence on whether calls move giving is honestly mixed, and it is worth knowing both sides before you invest volunteer hours. An observational analysis by donor software company Bloomerang of nearly two million first-time donors found that donors who received a thank-you call within 90 days were retained at higher rates than those who did not, with the effect growing when organizations called more than once. A large randomized field experiment published in the American Economic Journal: Applied Economics, however, found no measurable effect of thank-you calls on retention or gift size when the calls were made by an outsourced call center several months after the first gift. Bloomerang has argued that the timing and the outsourced callers explain the difference.

The practical lesson from both studies is the same: a call's value depends on who makes it, how soon, and how personal it feels. A prompt, warm call from someone connected to the mission seems far more likely to matter than a generic call months later from a stranger reading a script. That is the core of the argument in this article. AI cannot make a call feel personal by placing it. It can make a call feel personal by making sure the human caller knows who they are talking to, has something genuine to say, and follows up promptly afterward.

Building and Prioritizing the Call List

A calling program succeeds or fails on its list. A good list is short enough that callers can finish it, focused on the people most likely to welcome a call, and scrubbed of anyone who should not be called. Most organizations build lists by running a CRM query and exporting a spreadsheet, then hand-editing it. AI can make that process faster and more thoughtful, mainly by helping you reason about segments and catching problems in the export before a volunteer dials a wrong number.

Start by defining the goal of the campaign in a sentence, because the right list depends on it. A thank-a-thon after a spring appeal might target every first-time donor from the past sixty days plus every donor who increased their gift. A reactivation campaign might focus on people who gave in at least two of the past five years but not in the last eighteen months, which is a far warmer group than donors who gave once a decade ago. A monthly giving campaign might target donors who have given three or more single gifts in a year, since repeat donors are already behaving like sustainers. Our guide to reactivating lapsed donors goes deeper on recency and frequency segments, and our article on converting year-end donors to monthly giving covers the upgrade audience.

Once you have a draft export, a general-purpose AI assistant with spreadsheet capability can help in several concrete ways. It can propose a priority score that combines recency, frequency, gift size, and relationship notes, and explain how each factor is weighted so you can adjust it. It can split the list into balanced call sheets of a manageable size per caller, grouped by time zone so nobody calls a West Coast donor at 7 a.m. It can flag duplicate households, missing phone numbers, numbers in the wrong format, and records where the "do not solicit" or "do not call" field is set but the record was still exported. It can also suggest which donors should be routed to particular callers, such as giving long-time donors to board members and alumni donors to student callers from the same program.

The scrubbing step deserves extra care because it has legal and relationship consequences. Before any list goes to callers, remove everyone who has asked not to be called by your organization, everyone flagged as deceased, and anyone your gift officers have asked to manage personally. If a paid firm will make the calls, the federal rules described below require that donors who previously asked not to be called on behalf of your organization be excluded. AI can find the records; a person should confirm the final exclusions. Where possible, run the analysis inside the AI features your CRM already offers rather than pasting donor records into a consumer chatbot, a distinction we explore in our overview of AI-enabled nonprofit CRMs.

Scripts, Objection Handling, and Role-Play Practice

Good phone scripts are not read word for word. They are a structure: a warm opening that identifies the caller and the organization, a reason for the call, a moment to listen, an ask or a thank-you, and a clear close. New callers need more structure than experienced ones, and different campaigns need different scripts. AI is well suited to producing a first draft of each version and a set of variations, which a staff member then edits for voice and accuracy.

When you ask for a script, give the model real inputs: your mission in a sentence, the campaign goal, two or three concrete impact facts you can stand behind, the ask amounts or upgrade options, and the tone you want. Ask for a short version for confident callers and a fuller version for first-timers. For a monthly giving campaign, ask for language that explains how recurring gifts work, how to change or cancel them, and why predictable income matters to your programs, then verify every factual claim. Our guide to recurring donor programs covers the stewardship that should follow a successful upgrade call.

Objection handling is where volunteers most often freeze, and where AI preparation pays off. Ask the model to list the objections callers are likely to hear for this specific campaign, such as "I already gave this year," "I'm on a fixed income now," "How much of my gift goes to overhead?", "Take me off your list," or "Is this a scam?", and draft a short, respectful response to each. The best responses acknowledge the donor, avoid pressure, and offer a graceful exit. Some objections are not objections at all but instructions. "Take me off your list" and "don't call me again" are do-not-call requests, and your script should tell callers exactly how to record them rather than argue. "Is this a scam?" deserves a ready answer that tells the donor how to verify the call independently, for example by hanging up and calling the main number on your website.

Role-play practice is one of the most underused AI applications in fundraising training. A caller can open a chat or voice session with an AI assistant, describe the donor persona from a briefing, and practice the conversation: a delighted long-time donor, a donor who is mildly annoyed to be called during dinner, a donor who asks a pointed question about how funds are used. The AI plays the donor, and the caller practices in private, as many times as needed, without the awkwardness of rehearsing in front of colleagues. Afterward, the caller can ask for feedback on whether they listened, whether they made the ask clearly, and whether they handled the objection gracefully. This is a legitimate use of a synthetic voice, because nobody on the other end is a real donor. Use invented personas for practice rather than real donor records.

Role-play is especially useful for board members, who are often nervous about calls but rarely admit it. A fifteen-minute practice session before a board call night, with an AI playing three or four common donor types, can turn a reluctant trustee into a confident one. For student telefunds, where caller turnover is high, AI practice scenarios can supplement the supervisor coaching that already happens on the floor.

The Rules: TSR, TCPA, and AI-Generated Voices

Two federal regimes govern fundraising calls, and which parts apply depends on who is calling and how. The FTC's Telemarketing Sales Rule (TSR) governs telemarketing, and the FCC's rules under the Telephone Consumer Protection Act (TCPA) govern the use of autodialers and artificial or prerecorded voices. State laws add their own layers. What follows is an orientation, not a substitute for counsel, and the rules are detailed enough that any organization running a significant calling program should have them reviewed.

The Telemarketing Sales Rule and paid callers

Nonprofits calling on their own behalf, with their own staff and volunteers, are generally outside the FTC's jurisdiction and therefore outside the TSR. The picture changes when a for-profit firm makes the calls. Since the USA PATRIOT Act of 2001, the TSR has covered for-profit telemarketers soliciting charitable contributions on behalf of nonprofits, a practice often called telefunding. The FTC's guidance for telemarketers explains that these calls are exempt from the National Do Not Call Registry, but the telefunder must honor a donor's request not to receive calls on behalf of that particular charity. In practice, that means your organization and its telefunder need a shared, charity-specific do-not-call list that is updated promptly and applied before every campaign.

The rule's text, at 16 CFR 310.4, sets several other requirements for paid charitable calls. The caller must promptly disclose the identity of the charitable organization and that the purpose of the call is to solicit a charitable contribution. Calls to a person's home are limited to between 8 a.m. and 9 p.m. in the recipient's local time unless the person has consented otherwise. Prerecorded message calls are restricted; the rule permits them for charitable solicitations only to members of the nonprofit or previous donors, and only with specific opt-out mechanisms built into the message. The rule's definition of "previous donor" is narrow: someone who has given to that charity within the two years before the call. Misrepresentations about how donations will be used, or about the percentage of a gift that goes to the charity, are prohibited outright. Amendments that took effect in 2024 also expanded the TSR's recordkeeping requirements, including longer retention periods and call detail records that must identify the charity for which each call was made, so ask your vendor how it complies.

The TCPA, prerecorded voices, and the AI ruling

The TCPA applies to everyone, nonprofits included, and its central restrictions concern autodialed calls and calls using an artificial or prerecorded voice. For calls to cell phones, which now describes most donor numbers, a non-emergency call made with an artificial or prerecorded voice generally requires the recipient's prior express consent, and there is no general exemption for nonprofits. For calls to residential landlines, the FCC's rules allow a tax-exempt nonprofit to make a limited number of prerecorded calls without consent. Under rules that took effect in July 2023, those exempt calls are capped at three within any consecutive 30-day period, and each must include an opt-out mechanism, under the FCC's 2023 rule published in the Federal Register.

The FCC's February 2024 AI declaratory ruling made clear that AI-generated voices fall into the "artificial" voice category. The ruling did not create a new prohibition. It confirmed that an AI voice agent calling donors is subject to every restriction that already applied to prerecorded robocalls: consent requirements for calls to cell phones, the residential call limits, identification and disclosure requirements, and opt-out mechanisms. A conversational AI caller that sounds human does not escape these rules by being interactive. Combined with state laws, some of which are stricter than federal law, this makes AI-voiced outbound fundraising a significant legal risk for any organization without carefully documented consent. If you also run text campaigns, our article on text message fundraising and TCPA consent covers how consent works for texts, which the TCPA treats similarly to calls.

Calls placed manually by a human being, speaking in their own voice, sit in a very different position. A volunteer dialing from a call sheet and talking to a donor is not making a prerecorded or artificial voice call. That is one more reason the human-voice model described in this article is the safer design, though it does not remove the need to honor do-not-call requests, respect reasonable calling hours, and comply with any state rules that apply to your callers.

State registration for professional solicitors

Most states that regulate charitable solicitation also regulate the paid firms that solicit on charities' behalf, using terms such as professional solicitor, professional fundraiser, or commercial fundraiser. Requirements vary, but commonly include registration before soliciting, a surety bond, filing the contract with the charity or a notice before each campaign, required disclosures to donors during the call, and financial reporting after the campaign. California, for example, requires commercial fundraisers to register with the Attorney General and file a notice of intent before each solicitation campaign. Your organization usually has its own registration obligations as well, separate from the firm's. Our guide to charitable solicitation compliance with AI explains how to keep state registrations organized.

A Call Night Workflow, Start to Finish

Putting these pieces together, here is how an AI-supported board call night or thank-a-thon might run for a small development team. Two weeks out, the development director writes the campaign goal in a sentence and pulls a CRM export. An AI assistant helps segment and score the list, flags problem records, and splits it into call sheets by caller and time zone. A staff member confirms the exclusions and assigns sheets. One week out, the AI drafts briefings for every donor on the list, and staff spot-check a sample. The AI also drafts the script, the voicemail message, and the objection responses, which the director edits.

A few days before the event, callers receive their sheets, the script, and a link to an AI role-play exercise with three practice personas. Board members who have not called before are encouraged to practice at least once. On the night itself, nothing about the calls is automated. Callers dial from their own phones or a simple dialer, talk to donors in their own voices, and type brief notes into a shared form. A staff member is on hand to answer questions and to record any do-not-call requests immediately.

The next morning, the AI converts the raw notes into structured CRM rows and drafts follow-up letters and emails for every reached donor, every pledge, and every voicemail. Staff review do-not-call entries and pledges by hand, import the rest, and send the follow-ups within forty-eight hours. A week later, the AI summarizes the results for the board: how many donors were reached, what they said, what pledges came in, and what themes emerged from the conversations. That summary often turns out to be one of the most valuable outputs of the whole program, because it captures what donors actually think in their own words.