Pledge Drives and Underwriting: AI for Public Radio and Community TV
Local public radio stations, community radio, and community TV channels now raise most of their budgets from listeners, viewers, and local businesses. Federal support through the Corporation for Public Broadcasting is gone, and the remaining revenue streams have to work harder with the same small staff. This guide covers where AI can genuinely help: preparing pledge drive scripts and break plans, coordinating volunteer phone banks, growing sustaining memberships and winning back lapsed members, drafting underwriting copy that stays inside FCC noncommercial limits, and turning program logs into transcripts and reports. It also marks the places AI should stay out, including cloned host voices and anything that blurs the line between underwriters and the newsroom.
Anyone who has worked a public radio pledge drive knows the rhythm. The host comes out of the newscast, the producer slides a sheet of pitch points across the console, the phones light up or they do not, and someone in the back office tries to reconcile the online form, the phone bank tally sheets, and the thank-you gift inventory before the next break. Community TV and PEG access channels run a different schedule, but the pattern is familiar: a tiny staff, a large volunteer base, and a constant need to explain why local, noncommercial media deserves support.
The financial ground under that work moved sharply in the last eighteen months. In July 2025, Congress rescinded more than a billion dollars in previously approved funding for the Corporation for Public Broadcasting, and in January 2026 the CPB board voted to dissolve the organization. Many stations saw an outpouring of new members in response. Others, especially small rural stations that had relied on federal grants for a large share of their budgets, have already cut staff and programming. Either way, the money that keeps a station on the air now comes overwhelmingly from individuals, local businesses, and foundations.
That shift makes membership and underwriting operations the center of a station's survival, and it is precisely the kind of repetitive, deadline-driven writing and data work where AI tools earn their keep. A development director who once spent three days drafting pitch scripts can draft them in an afternoon and spend the rest of the time coaching hosts. An underwriting rep can produce a first draft of a sponsor credit in seconds, then spend real attention on making sure it complies with federal rules.
This article is an operations guide for US public radio stations, community and low-power FM stations, public TV stations, and community TV or PEG access channels. It walks through pledge drive preparation, volunteer phone banks, sustaining membership and lapsed-member win-back, underwriting sales under the FCC's noncommercial rules, program logs and transcripts, and the post-CPB funding picture. Throughout, it draws clear boundaries. The station's credibility depends on two things AI must never compromise: the honesty of what goes out over the air and the independence of the newsroom from the people who pay for underwriting.
The Funding Picture After CPB
For decades, Community Service Grants from the CPB provided a stable base for local stations, and for many small-market and tribal stations that base was a large share of the budget. With CPB gone, there is no federal successor program. Philanthropy has stepped in partly. In August 2025, a group of national foundations including Knight, MacArthur, and Ford committed $36.5 million in emergency funding, most of it routed through the Public Media Bridge Fund managed by Public Media Company. As the fund's name suggests, it is designed as a bridge to sustainability, not a permanent replacement. Eligibility rules and grant windows have changed between rounds, so check the fund's own site rather than relying on secondhand summaries.
Practically, this means three things for a station's operations. First, individual giving, and especially monthly sustaining giving, carries more weight than ever. Second, underwriting revenue matters more, which raises the stakes of getting FCC compliance right as sales teams push harder. Third, stations are writing more grant applications to foundations and community funders, each of which wants a clear story about local service, audience, and financial plans.
AI helps with all three, but the most immediate value is in planning. A finance lead can export three years of revenue by source, remove donor names, and ask an AI assistant to build scenarios: what happens if sustainer growth slows next year, if two major underwriters do not renew, or if a bridge grant ends on schedule. Our guide on surviving federal funding cuts walks through the broader method. Treat the output as a structured starting point for your board and finance committee, and check every formula by hand.
One caution on history. Stations that received CPB grants accepted certain obligations, and some reporting or recordkeeping tied to past grant periods may still matter for audits and closeout. Do not ask a chatbot what you still owe. Ask your auditor or counsel, and use AI only to organize the documents they request.
Planning Questions Worth Modeling
Use de-identified revenue data, then verify every number
- How many new sustaining members, at your average monthly gift, would replace the lost federal grant over twenty-four months?
- Which underwriters account for the largest share of underwriting revenue, and what happens if the top three do not renew?
- How much revenue did each pledge drive raise per on-air hour, and is a shorter drive with stronger sustainer conversion a better trade?
- What cash cushion do you need to absorb the gap between a bridge grant ending and new revenue arriving?
On-Air Pledge Drive Prep: Scripts, Breaks, and Gifts
Pitch Scripts That Sound Like Your Hosts
Good pledge pitches are specific. They name the local reporter who covered the school board, the volunteer who has run the Saturday folk show for twenty years, or the community TV crew that streamed every city council meeting last winter. Generic appeals about "the value of public media" are easy to write and easy to ignore. The work of a pledge drive producer is gathering those specifics and shaping them into short, conversational pitch points hosts can use in a ninety-second break.
AI is well suited to the shaping part. Give it your station's recent local coverage list, program schedule, a few listener comments you have permission to quote, and your drive goals, and ask for a pitch sheet organized by daypart: morning news, midday music, weekend specialty shows. Ask for several variations of each point so hosts are not repeating the same phrasing every hour. Then hand the draft to the hosts. Experienced hosts will rewrite most of it in their own voice, and that is the point. The tool saves the blank-page time; the host provides the warmth and credibility that actually moves listeners.
Two rules keep this honest. Never let AI invent a listener testimonial, a statistic about your audience, or a story about your coverage that did not happen. And never let it write claims about what the station will lose without federal funding unless your leadership has confirmed them. Pledge drives run on trust, and one inflated claim repeated every hour for ten days can do lasting damage.
Break Planning and Drive Calendars
Drive producers juggle dozens of variables: which programs perform best for pledges, which hosts are available, when matching gift challenges start and end, and when to switch from one thank-you gift to another. A general AI assistant can turn a messy spreadsheet of past drive results into a draft break schedule, highlight hours that historically underperform, and suggest where to place challenge grants for the most momentum. A human producer still makes the final call, since they know which host had a cold last spring and why Tuesday afternoon numbers were strange.
Thank-You Gift Logistics
Thank-you gifts, often called premiums, are a logistics problem dressed up as a fundraising tactic. Stations need to forecast quantities, track inventory, coordinate fulfillment, and handle backorders without upsetting members. They also carry tax obligations. When a member gives more than $75 and receives something of value in return, IRS rules generally require the station to provide a written statement estimating the value of the gift so the donor knows the deductible amount. IRS Publication 1771 explains these quid pro quo rules and the limited exceptions for low-cost token items.
AI can help you prepare: drafting the fair market value language for each gift tier, building a fulfillment checklist, forecasting quantities from past drives, and writing clear backorder notices. Our article on donor thank-you workflows covers the acknowledgment side in more depth. Have your finance lead or tax adviser confirm the value statements, since the thresholds for token items are adjusted for inflation each year.
Real-Time Pledge Tallies
During the drive, hosts want to know where they stand. Many stations already have dashboards in their membership software. Where they do not, a simple shared spreadsheet fed by online forms and phone bank entries can produce hourly totals, and AI can help write the formulas, flag duplicate entries, and summarize the last hour into two sentences a host can read on air. Keep the summary honest: if the tally includes estimates or pending card authorizations, say so internally, and only announce confirmed numbers on air.
Pledge Drive Prep Checklist
Where AI drafts and where people decide
- Gather real local stories, coverage examples, and permissioned listener quotes before drafting anything
- Ask AI for pitch point variations by daypart, then have each host rewrite them in their own voice
- Draft fair market value statements for every gift tier and have finance confirm them
- Build the break schedule from past drive results, with a producer making final calls
- Set up a tally process that separates confirmed gifts from pending ones
Coordinating Volunteer Phone Banks
Volunteer phone banks remain a fixture of public media drives, especially at community stations where answering phones in the studio is part of the culture. They are also a coordination headache. Shifts need filling, volunteers need a short briefing on current gifts and matching challenges, and the paper or digital pledge forms have to be complete enough for the membership team to process without calling the donor back.
AI fits naturally into the preparation. It can draft the shift sign-up messages and reminders, generate a one-page briefing sheet for each shift that reflects the current gift lineup and challenge totals, and produce a short answer guide for common caller questions: How do I change my monthly gift? Is my donation tax-deductible? Can I give in honor of someone? Our article on volunteer onboarding with AI describes how to build orientation materials that new volunteers actually read.
After each shift, AI can help clean up the data. A membership coordinator can paste in a batch of anonymized pledge entries and ask the tool to flag missing fields, inconsistent gift amounts, or likely duplicates. That catches problems while the caller's details are still fresh, rather than weeks later when a card fails. Keep full card numbers and other payment details out of any AI tool entirely. They belong only in your payment processor and membership system.
The calls themselves stay human. A volunteer answering a pledge line is often the only person a member will ever speak to at the station, and the tone of that conversation shapes whether they renew. Scripts should be guides, not word-for-word reads. If you are considering outbound calling campaigns, our guide to AI-assisted phone fundraising call sheets shows how to prepare callers without turning them into robots.
A Shift Briefing Sheet AI Can Draft
Regenerate it before each shift as gifts and challenges change
- Current thank-you gifts, giving levels, and what is backordered
- Active matching challenges and how much remains to unlock them
- A suggested sustainer ask for callers who make a one-time pledge
- Answers to the five most common caller questions
- Who to hand a call to when a member raises a complaint about coverage
Sustaining Members and Lapsed-Member Win-Back
Sustaining membership, meaning monthly recurring gifts, has become the backbone of public media fundraising. Sustainers renew automatically, give more over time, and reduce the pressure on every on-air drive. Many stations now treat converting one-time givers to sustainers as the primary goal of each drive, rather than the raw total. Our article on building recurring donor success covers the strategy in depth, and the conversion tactics in our guide to monthly giving conversion translate directly to member stations.
Keeping Sustainers From Slipping Away
The quietest threat to a sustainer program is involuntary churn: cards that expire, get replaced after a fraud alert, or decline for insufficient funds. Members often do not realize their support has stopped. AI can help you see this earlier. A membership manager can export a de-identified list of failed transactions with dates, amounts, and decline reasons, and ask an AI assistant to group them by pattern and suggest outreach timing. It can then draft a set of friendly, plain-language update requests by email and text, with a different tone for a first notice and a third.
Voluntary cancellations deserve a different response. When a sustainer cancels, the reason matters: a job loss, a move out of the listening area, or frustration with a program change. AI can help sort free-text cancellation reasons into categories so leadership sees trends. The response to an individual member, though, especially one upset about coverage, should come from a person.
Winning Back Lapsed Members
Every station carries a long list of former members who gave once or twice and drifted away. A thoughtful win-back campaign treats them differently by history: a former ten-year member who lapsed last year needs a different message than someone who gave once during a drive five years ago. AI can segment the list using giving history, draft tailored messages for each segment, and suggest which programs or local stories to mention based on the member's original giving context. The approach in our guide to re-engaging lapsed donors applies well here.
Be careful with outbound phone campaigns that use automated or synthetic voices. In February 2024, the FCC issued a declaratory ruling confirming that AI-generated voices count as "artificial" voices under the Telephone Consumer Protection Act, which brings them under the same consent rules as prerecorded calls. Nonprofit calls have some exemptions under the TCPA, but they are narrower than many people assume, and the rules differ for cell phones and landlines. Talk to counsel before any automated calling campaign, and keep live volunteer or staff calls as the default for win-back.
Good Fits for AI
- Grouping failed payments by decline reason
- Drafting card update reminders in a warm tone
- Segmenting lapsed members by giving history
- Categorizing cancellation reasons for leadership
- Writing upgrade asks for long-time sustainers
Keep With People
- Replies to members upset about coverage or programming
- Conversations with major and planned giving prospects
- Outbound calls, unless counsel has cleared automation
- Handling payment details and refunds
- Any decision to remove or suppress a member record
Underwriting Sales Inside the FCC's Noncommercial Limits
Underwriting is where AI can save the most time and also create the most risk. Noncommercial educational (NCE) stations, including public radio and TV stations and community and low-power FM stations, cannot air advertisements under Section 399B of the Communications Act and the FCC's rules. They can air acknowledgments that identify a sponsor. Since the 1980s, the FCC has allowed so-called enhanced underwriting, which can include a sponsor's name, location, logo or established slogan that identifies rather than promotes, a value-neutral description of its products or services, and contact information such as a phone number or website.
What enhanced underwriting cannot contain is spelled out in decades of FCC enforcement. Credits may not include calls to action ("visit today," "call now," even "planning a special occasion?"), qualitative or comparative claims ("the finest," "award-winning service," "the area's best"), price or discount information, or inducements to buy. The FCC expects stations to make good-faith judgments about each credit and has issued fines against stations whose credits drifted into promotional language. The FCC's 1986 policy statement on underwriting remains a foundational reference, and the Student Press Law Center's underwriting guide offers a readable summary. Your communications counsel is the final authority.
Here is the problem with AI. General-purpose writing tools have been trained on enormous volumes of commercial advertising copy. Ask one to "write a fifteen-second sponsor message for a local bakery" and it will almost certainly produce something like "Treat yourself to the best croissants in town, stop by today!" That single sentence contains a qualitative claim and a call to action. Left unchecked, AI drafting will pull underwriting copy toward advertising every time.
The fix is a combination of strong instructions and mandatory human review. Write a standing prompt, or a saved custom instruction, that spells out the station's underwriting policy in plain terms, lists prohibited words and constructions, and includes several approved credits as examples. Ask the tool to produce neutral copy in your station's format, and then ask it to check its own draft against the policy, listing any words that might be qualitative or promotional. That self-check catches many problems. It does not replace a trained reviewer, who must approve every credit before it airs.
AI can also help with the sales process around the copy. It can research local businesses for prospect lists, draft proposal letters that explain the audience and the noncommercial format, and build renewal reminders. The principles in our guide to corporate sponsorship packages apply, with one difference: underwriting proposals must never promise what the station cannot deliver under FCC rules, such as on-air endorsements by hosts or coverage of the sponsor.
Generally Permitted in Credits
Confirm against your station's own policy
- Sponsor name, location, and years in business
- Value-neutral description of products or services
- Established slogans that identify rather than promote
- Website address and phone number
Not Permitted
The words AI drafts tend to add
- Calls to action like "visit," "call," "stop by," or "try"
- Qualitative words like "best," "finest," "delicious," or "trusted"
- Prices, discounts, sales, or free offers
- Comparisons with competitors or inducements to buy
A Human Reviews Every Spot
No AI-drafted credit should reach air without approval from a trained staff member who knows the station's underwriting policy. Keep a record of who approved each credit and when. If a sponsor pushes back on neutral language, the answer comes from a person, not a revised prompt.
Community TV and PEG access channels are not FCC broadcast licensees, so Section 399B does not govern them directly. Their sponsorship rules come from local franchise agreements, municipal policies, and the channel's own operating rules, and many prohibit commercial content outright. Give the AI tool your actual local policy, not a general description, and have your channel manager approve every message.
Protecting the Firewall Between Underwriters and the Newsroom
Public media's credibility rests on the understanding that underwriters do not shape coverage. Most stations maintain a firewall: sales staff do not discuss story plans with sponsors, reporters are not told who is underwriting what, and a sponsor's support never earns favorable treatment. AI introduces new and subtle ways for that firewall to leak, and stations should anticipate them.
The simplest leak is shared tools and shared data. If the development team's AI workspace contains underwriter contracts and the newsroom uses the same workspace with shared project files, a reporter's research query could surface sponsor information. Keep separate accounts or projects for development and editorial work, and be deliberate about what each can see. The same applies to CRM systems with built-in AI features: if your membership database can summarize a person's relationship with the station, make sure reporters are not using that feature to look up sources.
A subtler risk is AI-assisted prospecting that targets businesses based on coverage. A sales rep who asks a tool to "find local companies featured in our recent stories" and then pitches them underwriting creates exactly the appearance of a connection the firewall exists to prevent. Write your prospecting rules down, and include them in your AI use policy. Our article on AI in nonprofit newsrooms covers editorial AI standards in more detail, and NPR's ethics handbook section on generative AI is a useful model for local policies.
Firewall Rules for an AI Era
- Separate AI workspaces and shared files for development and editorial teams
- No underwriting prospect lists built from recent coverage
- No AI-generated "sponsored content" that resembles news or program segments
- Newsroom access to CRM AI features limited or disabled
- Firewall rules written into the station's AI use policy and reviewed by the board
Program Logs, Transcripts, and Reporting
Behind every station is a pile of documentation: program logs, underwriting airing records for invoicing, and, for broadcast licensees, the quarterly issues and programs lists that go in the FCC public inspection file. These lists describe the local issues the station covered and the programs that addressed them. They are tedious to compile and easy to fall behind on, and they matter at license renewal.
AI transcription has made this much easier. A station can transcribe local programs and news segments, then ask an AI tool to summarize each segment and tag the community issues it addressed, such as housing, schools, public safety, or the local economy. A staff member then reviews the draft list, corrects any mislabeled segments, and approves it for the public file. The tool does the sorting; a person signs off on what the station tells the FCC it did.
Transcripts have value well beyond compliance. They make audio searchable for the newsroom, support accessibility for deaf and hard-of-hearing audiences, and give the development team accurate quotes for grant applications and donor reports. Our guide to AI for captions and transcripts covers accuracy checks and workflows. For community TV channels, AI-assisted captioning of city council meetings and local programming can be a meaningful public service, though names, places, and technical terms always need a human pass.
Underwriting reconciliation is another strong use. Comparing contracted credits against the automation system's airing log by hand takes hours. A spreadsheet export and a well-structured prompt can flag missed airings, overruns, and credits that ran in the wrong daypart, so the traffic manager can schedule make-goods before invoices go out. As always, the airing log itself remains the record, and any discrepancy is confirmed by a person.
Documentation Tasks Worth Automating First
- Draft quarterly issues and programs lists from transcripts, reviewed before filing
- Reconcile contracted underwriting credits against airing logs
- Produce searchable transcripts of local programs for the archive
- Draft local service summaries for foundation grant applications
- Caption recorded public meetings on community TV channels
Where AI Does Not Belong in Public Media
Some uses of AI are not just risky; they cut against what makes public media worth supporting. The relationship between a host and a listener is personal. People give because they trust the voice they hear every morning, and they expect that voice to be real.
Cloning a host's voice for pledge breaks, station IDs, or promos without explicit, informed, written consent is a line no station should cross. Even with consent, think carefully. Listeners who later learn that a heartfelt pledge appeal was synthesized may feel deceived, and the damage spreads beyond that one host. If a station ever uses a synthetic voice for routine tasks like weather or overnight IDs, disclose it clearly and keep it far away from fundraising appeals and news. Volunteer hosts deserve the same protection as paid staff, and agreements should say plainly who controls a host's voice and likeness.
Other boundaries are just as firm. Do not use AI to fabricate listener testimonials, generate fake call-in voices, or create images of community members for fundraising materials. Do not let AI write news copy that reaches air without editorial review. Do not use AI to decide which members get thank-you gifts or service based on predicted wealth in ways your members would find intrusive. And do not let a sales team use AI to tailor underwriting pitches around editorial content. Each of these trades a small efficiency for the trust that makes a station worth funding.
The Boundary in One Sentence
AI can draft, sort, and summarize the work around the broadcast, but every voice, claim, and credit that reaches the audience is real, verified, and approved by a person.
A Practical First Ninety Days
Most stations have no time for a technology project, and they do not need one. Start with the tasks that cost staff the most hours and carry the least risk, write down the rules before anyone starts, and build trust gradually. A good sequence ties the work to your next on-air drive so the payoff is visible.
Days 1-30
Set the rules
Write a one-page AI policy covering member data, underwriting review, voice consent, and the newsroom firewall. Build your underwriting prompt with approved examples. Model revenue scenarios for the board.
Days 31-60
Fix the membership work
Launch failed-payment outreach for sustainers. Segment lapsed members and draft win-back messages. Prepare gift value statements and volunteer briefing templates for the next drive.
Days 61-90
Run the drive
Use AI-drafted pitch sheets that hosts rewrite, shift briefings for phone volunteers, and hourly tally summaries. Afterward, compare sustainer conversion and staff hours to the prior drive.
Conclusion
Local public media is in a harder position than it has been in decades. With the CPB dissolved and philanthropy offering a bridge rather than a replacement, stations depend almost entirely on members, underwriters, and local funders. AI will not close that gap on its own. It can, however, give a small development team back the hours it spends on drafting, sorting, and reconciling, so people can spend more time with members, sponsors, and the community.
The best uses are practical and bounded: pitch sheets that hosts make their own, volunteer briefings that stay current, sustainer outreach that catches failed cards early, underwriting drafts that start neutral and are approved by a trained reviewer, and transcripts that turn hours of programming into searchable records. Each keeps a person accountable for what goes out over the air.
The limits matter just as much. Keep host voices real, keep testimonials true, keep underwriting credits inside FCC rules, and keep the newsroom independent of the people who pay for credits. Listeners and viewers support public media because they trust it. Every AI decision a station makes should protect that trust first.
Make Your Next Pledge Drive Easier to Run
We help public radio stations, community TV channels, and community media nonprofits put AI to work on membership, underwriting, and reporting, with clear policies that protect your newsroom and your listeners' trust.