Inside the Race to Build AI Bots for Fundraisers
By Laurence A. Pagnoni, MPA
Author, The Nonprofit Fundraising Solution | Former Faculty, NYU Heyman Center for Philanthropy, School of Professional Studies
Published July 2026 | laurencepagnoni.com
Exec Summary
Lawyers have the AI bot Harvey — an AI tool built specifically for the legal profession that's become nearly indispensable at top firms in just a few years. We fundraisers don't have an equivalent yet. What we have instead is a dispersed, fast-growing field of AI tools for fundraisers, each solving a piece of the donor-relationship puzzle, plus a genuine debate underway about what these tools should, and shouldn't, be allowed to do. I walk you through what AI tools already exist, how you might choose between them, what fundraisers using these tools are actually saying, one prominent voice raising an alarm about where this could go, and whether the pieces already exist for you to build a custom AI suite of your own. I also raise the question whether we even need our own version of Harvey.
Why Law Got Harvey
Harvey was built to solve an issue that was clear, expensive, and repetitive: contract review, legal research, and case analysis eat enormous billable hours, and clients were happy to pay for a tool that did it faster. That challenge was well-defined enough that a single AI product could take it on directly and it worked well enough that Harvey is now valued at $8 billion, with the majority of the top ten U.S. law firms among its customers. (Source: Fortune)
Our challenge isn't the same because our issues are not primarily a document-throughput dilemma waiting to be automated, but, more accurately, we have donor relationship dilemmas. Our real need isn't processing more paperwork faster; it's authentically customizing the donor experience to better understand what a specific donor cares about, where they are in their relationship with your organization, and what a meaningful next conversation with them actually looks like. That's harder for a single tool to "solve" the way Harvey solved contract review, and it may be why a category leader has not yet emerged for us.
What's emerged instead is a batch of separate tools, each one handling a different piece of that relationship.
AI Tools for Fundraisers: Who's Building What
Practivated, built by longtime nonprofit consultant Mallory Erickson, created Tivy — an AI donor coach that lets you rehearse a donor conversation before it happens, whether that's making an ask, handling an objection, or following up after a decline. (Source: DonorPerfect) GoFundMe took a different angle with its Smart Fundraising Coach, built for individual and peer-to-peer organizers rather than institutional shops, helping people craft their story and reach the right donors using patterns drawn from millions of past campaigns. (Source: Business Wire)
Bonterra's Network for Good combines your donor database with built-in coaching and best-practice recommendations, aimed squarely at small nonprofits without a dedicated development strategist on staff. (Source: Bonterra). Another platform, Golden, took the approach of donor-facing AI with Goldie, a coaching companion built into its volunteer-management platform that interacts directly with volunteers and donors, aiming to help each one grow into a more engaged supporter over time. (Source: SSIR) And then there is DonorSearch Ai and Dataro. They focus on analyzing a donor’s past giving and wealth data to tell you who’s likely ready for a bigger ask, and roughly how much to request. In the jargon of the AI filed, they “lean predictive.” But users will have to assess if their prospect research and wealth data is compelling. (Source: DonorSearch)
A smaller, more personal category is also emerging: individual consultants training their own AI on their books and client materials, then offering it to the rest of us. Fundraising consultant Lauren Brownstein's "Your Fundraising Coach" is one example, built to answer stewardship and board-communication questions in her own voice. (Source: PITCH Consulting)
And then there's Aardie, from Project Aardvark — a free demonstration built by legendary Steve Nill, founder of CharityChannel, the nonprofit sector's largest professional community. You can try it yourself at charitychannel.com/aardvark. Aardie is meant to show you what an AI guided by an explicit, published ethical framework looks like in practice — more on that shortly.
None of these has Harvey's scale or brand recognition. This is still a field of “point solutions,” in the jargon of the AI field, without a category leader emerging; but we face the question of whether we need one!
How to Choose the Best AI Fundraising Tool
With no clear category leader, choosing comes down to what dilemma you're actually trying to solve, not which tool has the most buzz.
If your biggest issue is confidence walking into a donor meeting, Tivy's rehearsal model is built for exactly that. If you're running a peer-to-peer or one-person campaign rather than an institutional shop, GoFundMe's coach is tuned to that context, not major-gift strategy. If you're a small nonprofit without a dedicated development strategist, Network for Good folds coaching directly into your CRM so you're not managing a separate tool at all. If your team already leans on volunteers as a donor pipeline, Goldie's volunteer-facing design may fit better than a standalone donor-coaching app. If what you actually need is a sharper list of who to ask and how much, DonorSearch Ai or Dataro will do more for you than a conversational coach ever could. And if what you want isn't tactical help so much as a second opinion grounded in an explicit ethical stance, for example a check on whether an ask makes sense given where the real donor relationship actually stands, not just how long it’s been since your last visit, then Aardie is worth trying, free, precisely because it's built to slow you down rather than push you toward the bigger number.
Overall it’s best to pick the AI tool that matches the specific moment in your donor relationship where you need support, not the one with the flashiest demo.
What Fundraisers Are Saying About AI Coaching
Vendor claims are easy to find, but hearing from people actually using these tools is useful.
At UnityPoint Health, a fundraiser used Practivated to prepare for a meeting where she anticipated a donor objection she hadn't planned for. She ran the exact scenario through the tool, debriefed with its AI coach afterward, and went into the real conversation better prepared. (Source: Practivated) Jake Heuser, VP of System Philanthropy at UnityPoint Health, who oversees fundraising across a dozen foundations, pointed to her as an example of pairing the right tool with the right person — and named a pattern he sees across our field: leaders often rush into prospect strategy without enough grounded preparation first. (Source: Practivated)
On the peer-to-peer side, GoFundMe organizer Windy Pham described her platform's AI coach in practical terms: it takes years of hard-won trial-and-error lessons and turns them into clear, real-time guidance. (Source: Business Wire)
Both accounts describe the AI as preparation used before or after the donor conversation, never as a stand-in for yourself.
The Fork in the Road: Coaching You, or Replacing You?
The distinction that AI prepares you versus AI talking to the donor directly is Steve Nill’s main concern.
Nill has spent more than two decades advising nonprofits and has helped build fundraising programs that produced nearly $1 billion in charitable contributions. (Source: Philanthropy News Digest) He's straightforward about where he thinks our industry is headed:
"My fear is the trend of software vendors inserting AI between the fund development professional and the donor — some CRM providers are even encouraging a reduction of development staff in favor of donor-facing AI. This isn't just about protecting jobs; it's about making sure AI stays beside the fundraiser, strengthening the human relationship with donors rather than replacing it." — Steve Nill, founder, CharityChannel
Nill's answer was to write a book on the subject, Holding Fire: A Skeptic's Framework for the Fundraiser with AI at Your Side, and to publish a companion set of ground rules called the Human-Centered AI Framework Operating Specification under a public license, in hopes it gets adopted by the major nonprofit CRM providers.
To see what that looks like in practice, I asked Aardie a real fundraising question: how do you break out of a private foundation's habitual $10,000 gift and ask for a larger gift? Rather than handing back a script or a generic ask-escalation formula, Aardie started by naming what my own donor records didn't yet show — whether $10,000 reflected the foundation's ceiling, its comfort level, or simply its standard giving formula. It then reframed the goal entirely: instead of pushing for a bigger number, Aardie suggested that I invest in a real conversation with the program officer to learn their constraints and interests first, and treat a larger ask as something that follows from that relationship rather than something forced ahead of it. It even flagged the risk of getting this wrong which was that chasing a bigger gift before understanding a donor's capacity can read as chasing their money rather than serving the mission.
That's a useful, concrete illustration of the distinction Nill is drawing. It's coaching that pushes you toward a stronger relationship, not a script that replaces one.
How to Build Your Own AI Fundraising Suite Today
Setting the ethical debate aside, is any of this buildable in-house?
Increasingly, yes. Anthropic who is the maker of the Claude AI models offers Claude for Nonprofits, a program built for mission-driven teams that connects directly to the donor-data infrastructure you already use: Blackbaud for CRM and campaign tracking, Benevity for donation and volunteering data, and Candid for grants and funder research, all at discounted nonprofit pricing. (Source: Tygart Media)
On top of that base, Anthropic's "Agent Skills" let you build small, reusable instructions Claude can follow. For example, you can draft stewardship emails, one for prepping major-gift talking points, one for summarizing a donor's giving history that stack together into something close to a custom fundraising suite. (Source: Substack)
Making your Agent Skills real doesn't take an engineering team. Here's the steps to create them:
Start with Claude for Nonprofits, not a generic Claude account, so your donor-data connectors and nonprofit pricing are already in place.
Pick one recurring task first, not the whole suite at once, for example pick stewardship emails, or a talking-points prep sheet for major-gift meetings, or a donor-history summary. These are all good starting points because you already know what "good" looks like for each.
Write down how you do that task well today, for example note the questions you ask yourself, the tone you use, the things you always check before hitting send. That's the actual content of a Skill; it's less code than it sounds like, closer to writing down your own best practices in plain language.
Turn that into a Skill file Claude loads automatically whenever that task comes up, so you're not re-explaining your approach in every conversation. I created a WORD guidance document that I upload as needed to inform Claude of my approach.
Test it on a real, low-stakes case before trusting it on a major donor, and adjust the instructions based on where it gets the tone or judgment wrong.
Add the next Skill only once the first one is reliable.
Of note: a community-built starter kit already covers grant writing, donor communications, and volunteer coordination, free to install, and is a reasonable place to borrow from rather than starting from a blank page. (Source: Nonprofit Skills)
So far, nobody has assembled these pieces into a single, polished tool built specifically for us — whether that's a gap waiting to be filled, or simply evidence that our profession doesn't need one dominant tool the way law did. The raw materials exist either way: the fundraiser willing to build one Skill at a time may end up further ahead than the one waiting for a vendor to build something more comprehensive.
Where AI Fundraising Tools Go Next
Maybe that's the right outcome. If our real need is a genuinely customized donor experience rather than a repetitive process to automate, then a dispersed field of focused tools may serve us better than one dominant bot trying to do everything. Or maybe a category leader is still coming and the real question is whether it gets built with Nill's human-centered framework baked in, or without it.
Either way, the tools to build it already exist. Who builds next, and what they optimize for, will decide which future we will have.
A Note on Use
This white paper is offered freely for educational purposes. Please share it with executive directors, development directors, board members, and operations staff who may find it useful — provided the author's byline remains intact: By Laurence A. Pagnoni, MPA. Reproduction in publications, training programs, or institutional materials requires attribution. Produced July 2026 — to be reviewed and updated annually. All fees and platform details should be verified directly with each platform before any decision is made. Your comments are welcome in the comments section of the website.
