On March 19th, Art Weinkofsky (Director of Development Information Systems, Central Park Conservancy), Emily Palmer (Associate Director of Fundraising Data Management, Share Our Strength), and Emma Zawacki (Senior Development Operations Coordinator, Center for Reproductive Rights) joined us for an in-depth conversation on a function that is overworked, underappreciated, and — it turns out — one of the best candidates for AI-powered efficiency in the nonprofit sector.
Here's what they shared.
5 Key Takeaways
1. Gift processing is the engine that makes fundraising run, and it’s often overlooked
"It's the ultimate in sausage making. Nobody wants to know how it's made. They just want the end product."
What that sausage-making actually involves: logging into dozens of different portals, manually reconciling gifts with donation data, many of which arrive without clear information about who sent them or why. It's a puzzle, and when it's done well, it's invisible. When it's done poorly, it disrupts every part of the fundraising team.
%2520(1).png%3Fv%3D2026-03-19T22%253A15%253A26.620Z&w=2048&q=85&dpl=dpl_6CPVZ5cfhiV5uhKV4brjpEwEwhJr)
2. The job is only getting harder
At the Nonprofit Technology Conference (NTC) in Detroit last week, every attendee at a Gift Processing session said gift processing has gotten harder over the last five years.
The reasons: DAF giving volume is up, and the number of sources is multiplying. According to the DAF Research Collaborative’s Annual DAF Report, grantmaking from DAFs increased 19.0% YoY to $65 billion in 2024. The DAF Fundraising Report shares a similar trend, with 30% YoY DAF revenue growth across participating nonprofits.
And between DAF providers and workplace giving platforms, the number of DAF sources continues to grow.
"We have a list, literally a spreadsheet of all of our third party platforms. Last I checked, it was well over two hundred logins."
Managing 200+ portals — many with multi-factor authentication, some tied to personal phone numbers, some whose credentials were last updated by a volunteer from twenty years ago — is not a staffing problem. It's a structural one.
One of the largest nonprofits in the country recently went so far as to tell workplace giving platforms they wouldn't accept employee matches because they couldn't justify the processing burden. That's the kind of impossible trade-off that happens when volume outpaces capacity.
3. AI is an untapped opportunity in gift processing
At that same NTC session, the room was asked who was using AI for gift processing. Not a single hand went up.
"Gift processing is both complex, as we saw, and incredibly rote as well. And that kind of combination of roteness and complexity can lead to a lot of human error — but also burnout and morale loss... Any kind of place where repetition and complexity are connected so deeply, AI can just really be super helpful."
4. AI needs to be used responsibly
Art put it directly:
"We look at AI as an assist. It's not doing the work. It's a trust and verify type system."
Central Park Conservancy follows established internal policies, sets clear limits on what data can be used and how, and treats AI the same way they treat any other technology partner: with scrutiny around whether the data stays contained.
Chariot's gift processing software operates as a closed loop—its data is not used to train any external LLMs and data from Chariot’s customers is restricted to their individual account.
5. AI helps with the most tedious parts of the job
Art's team at Central Park Conservancy runs on a small-but-mighty model. "We've got two full-time people on our gift processing staff and we had to handle forty-six thousand gifts last year." The reaction when AI-assisted coding started working wasn't fear — it was relief.
"No one's feeling like anything's being taken away. If anything, they're feeling like they've got the better shovel... to dig out from under the load."
What You Can Try: Practical Tips for Building Coding Rules
The panelists were specific about how they got started. These are the steps that worked across three different organizations.
1. Start by segmenting your gifts BEFORE you write a single rule
Emma's first move was dividing gifts into groups that came in similarly and needed to be reported on similarly: DAFpay gifts, regular DAF gifts, and workplace matching gifts. Each follows different logic. Trying to write universal rules before doing this segmentation creates more problems than it solves.
2. Map which fields can be assigned immediately versus which depend on other fields
Some fields are straightforward; others build off each other. At the Center for Reproductive Rights, fund ID was dependent on constituency, which was derived from donor ID — meaning the rule had to account for that chain. Charting these dependencies before building rules saves significant rework later.
3. Feed your existing reference materials directly into the system
1.png%3Fv%3D2026-03-19T22%253A22%253A56.933Z&w=2048&q=85&dpl=dpl_6CPVZ5cfhiV5uhKV4brjpEwEwhJr)
Art's team manages roughly a dozen programs, each with its own set of appeals, fund codes, and letter codes for acknowledgements. Emma and Emily mentioned having thousands of potential codes to manage. That full list — what used to require manual lookup for every gift — now lives in the platform. The AI applies it automatically based on the rules the team has set. You can also just describe rules in plain English through the chat interface and the system will generate and apply them. Either way works.
4. Build rules for the platforms that give you the most trouble
Share Our Strength receives both regular DAF gifts and workplace giving gifts through Groundswell — and codes them differently.
The rule: "If it says 'employee match' in the note of the gift, code it as a workplace gift."
That single rule eliminates a manual classification step on every Groundswell import. Think about which portals cause your team the most friction and start there.
5. Let purpose-field logic do more work
Art's team trained the AI to flag keywords in gift purpose fields — words like "women's committee," "playground partners," "benches," "trees" — and use those to assign classification.
A $25,000 gift automatically routes to the major giving category.
1.png%3Fv%3D2026-03-19T22%253A24%253A07.406Z&w=2048&q=85&dpl=dpl_6CPVZ5cfhiV5uhKV4brjpEwEwhJr)
A gift mentioning a specific event gets the event's fund ID. Gifts that mention an event that was already in the system get matched to it.
The AI is reading the same cues a skilled gift processor would read manually.
6. Use flags for the things that still need a human
Emma set up a rule so that any gift with a purpose mentioning "in memory of" or "in honor of" generates a small exclamation point in an alert column. She doesn't have to scroll through hundreds of gifts to find the ones that need extra attention — the system surfaces them.
The same logic applies to anything you want a human to review before finalizing: restriction flags, address mismatches, unusual gift amounts. Build the rule, let the AI flag it, and move on.
7. The Benevity file can go from eight hours to thirty minutes
Emily's team used to put Benevity gifts in as a lump sum donor record. Now, with coding rules in place, the monthly file comes in fully coded to the individual employee and employer level — hard credits, soft credits, aliases for matching — in roughly half an hour.
"In the past, if we had wanted to get that employee-level data, we're looking at six to eight hours of work."
The process: export your workplace-giving company list with IDs from your CRM, upload it to the platform, and let the system match company names on incoming Benevity files to your reference list. When names don't match exactly, flag them for manual review.
A bonus: individual employee data often includes email addresses. Those are warm leads who've already given. Share Our Strength is now building stewardship relationships with donors they previously had no direct contact with at all.
8. For check processing, consider a digital mailbox
Central Park Conservancy launched Chariot's digital mailbox integration in November 2025, starting with checks mailed by DAFgiving360. They've since eliminated 500+ checks annually from manual processing — and are planning to expand to third-party bill pay services (Citi and others), which they estimate will remove another 500+ checks per year from the queue.
The digital mailbox deposits the funds, pulls the relevant data from each check and grant letter image and matches it to the standard data structure. Every gift through the Digital Mailbox appears in the Chariot dashboard the same as electronic gifts, and can have the same coding rules applied. The digitally-scanned check and grant letter images are also available if anyone needs to verify.
What This Actually Unlocks
The time savings are massive, but they're not the point on their own.
Emily put it simply:
"There is something to be said for working with a tool that was built for your day-to-day life... So often in operations, we are working with tools that were really made because the fundraisers needed them. The Chariot gift processing platform is a platform that was made for what we do on a day-to-day basis... Someone finally saw me and what I deal with on a day-to-day basis."
When gift processing teams spend less time on manual lookup and more time on the exceptions that actually need a human, the quality of data going into the CRM improves. Donor stewardship gets sharper. Fundraisers get better information. The function itself starts to look like what it is: a site of meaningful innovation, not a cost center.
Chariot gift processing combines a financial account with AI-powered data transformation to solve a fast-growing operational challenge that costs the 1.4 million nonprofits and thousands of grantmakers in the U.S. millions of hours annually. If you’re interested in learning more, get in touch with our team.
Coming Soon: A Monthly Gift Processing User Forum
One of the most energizing moments of this conversation happened in real time: Emma described her honor/in-memory flag policy, and Art immediately wanted to implement it. Emily walked through the Groundswell rule, and Art started connecting it to his own Benevity setup. The learning applications were immediate, and it surfaced a gap.
Gift processing professionals don't have dedicated spaces to share what they're building with each other. Starting soon, Chariot will be hosting a monthly gift processing user forum — open to anyone using or curious about Chariot’s gift processing platform. It will be focused on sharing AI-assisted policies, surfacing use cases, and learning from what's working working for your peers.
