So you’re done with pilots and you’re ready to implement an AI platform at your nonprofit. Which one do you choose? This is a tough decision facing all businesses these days and social good organizations have a couple of additional layers of complexity to think through before making a call.
As you plan your technology strategy for the new year, I wanted to explain how we approach this with our clients. I hope it helps you feel more confident with your decision.
Big AI vendor comparison for nonprofits
First, let’s look at the tools themselves. You might be tempted to make a decision solely on price but we caution against it. Remember the nonprofit technology adage “free as in kittens.” There are always tradeoffs to be had. Our job is to help you identify and navigate the ones that are important to your organization. Here is a comparison based on our experience with AI tools.
Core AI offer to nonprofits
Nonprofits can access the medium-weight model Sonnet as well as Haiku, which is great for quick work. Enterprise organizations can request access to Opus, a powerful model with advanced skill and agentic capabilities. Claude users can leverage the browser chat as well as voice mode (conversational AI), and Cowork, which can run tasks inside your apps and take over entire intelligent workflows.
Nonprofit packages and discounts
Most nonprofits are offered a discount of up to 75 percent. In our comparison of tools for a 120-member organization, Claude came in at $14,400 a year—half the cost of ChatGPT. The pricing model is a flat subscription tier but heavy API usage and Claude Code are charged separately and depend on the model used (Haiku outputs are around $5 per million tokens vs Fable $50 per million).
Advantages and Tradeoffs
There’s a reason most heavy users prefer Claude: Its writing and reasoning quality is generally better, out of the box, and it learns the how and why of your organization quickly. For generative writing like grant applications and board reports, there’s no better tool. Claude has 350 native connectors and an emphasis on safety and reasoning, which helps protect nonprofit reputation and donor data.
Anthropic is also the only vendor who is leaning into growing the nonprofit AI sector, with the vaunted Claude Corps fellowship, free courses, a partnership with GivingTuesday, and strong connections with implementation partners.
Core AI offer to nonprofits
ChatGPT is leading in brand recognition and user base and it’s likely some of your team already use it at home. Nonprofits have access to ChatGPT Business (self-serve billing and setup) and Enterprise (starting at 150 seats). The main differences are in whole-team features, rather than in the platform itself: Enterprise has SSO, SCIM user provisioning, and customized data retention policies. It also gives you access to dedicated account support. ChatGPT also has desktop, browser, and in-app workflows, and a conversational AI option.
Nonprofit packages and discounts
OpenAI doesn’t release its rates publicly but we were quoted a starting rate of $28,800 a year for 120 seats, which was a 20 percent discount. Other nonprofits have secured a discount of up to 75 percent. Like Claude, ChatGPT also has fees per seat for most use but charges more for advanced data analysis, custom agents, and deep API calls, dependent on model ($1.20 per million for Luna and $30 per million for its flagship model Sol).
Advantages and Tradeoffs
ChatGPT’s rack rate is the most expensive but it’s often heavily discounted. It has lots of built-in agent development tools and integrates with 1,400 other platforms. OpenAI tends to move faster with new capabilities, releasing bug fixes and improvements weekly and entirely new models every 30 to 90 days! If you need to be on the cutting edge or want to build an army of agents, this could be a good choice.
Core AI offer to nonprofits
Nonprofits can access Copilot (the full platform, which can operate from the desktop and in Microsoft apps) and Copilot Chat (just the assistant interface). Its workflow options are more tied to tiers, with the coworking and in-app AI flows only available to enterprise customers.
Nonprofit packages and discounts
For nonprofits already in the Microsoft ecosystem, Copilot can be a lot more affordable because it’s billed as an add-on rather than a whole new tool (And you can usually get it for a 15 per cent discount.) Kudos to Microsoft for making its cost structure visible: You can ballpark the costs for your organization’s needs at the Microsoft Nonprofits portal. Microsoft has a hybrid pricing model comprising a foundational Microsoft 365 license + a per-seat fee for Copilot + a metered AI credit consumption level (about 1c per credit). Credits work differently to tokens and they’re tied to feature use, as opposed to models.
Advantages and Tradeoffs
Copilot Chat is available on every Microsoft 365 tenant so you can try it out for free. It is definitely a little brother to the two big players, with 100 data connectors and fewer agentic capabilities. However, it does seamlessly integrate with Microsoft products and SharePoint, which makes sense because it’s built on the same stack. It has a fully built-out scenario library for nonprofit workstreams like grant tracking, funding research, and donor pipeline management.
Core AI offer to nonprofits
The chat version of Gemini operates inside Google Workspace and has a free tier for up to 2,000 users. Its advanced tier has coworking capabilities, a research engine, and built-in AI image and video generation features.
Nonprofit packages and discounts
If you work in Google Drive, you likely already have Gemini’s free tier all up in your business. The enterprise plan is among the cheapest—as low as $3.50 per seat per month for nonprofits—and for that, you get a 1 million-token context window, 2TB of storage, and the coworking functions that means Gemini can help you make sense of that horrible spreadsheet you’ve been ignoring for days. Additional tokens vary from $2.50 per million for Flash to $18 per mission for the Pro model in its highest context mode.
Advantages and Tradeoffs
Google’s model is different from the other players, and the bundled nature of it can be comforting to nonprofits. However useful Gemini is for research and coworking, its agent capabilities (called Gems) are less powerful than others in the sector. It is a good option for Google-ensconced nonprofits who want to be more efficient without fundamentally changing how they work.
How do AIs connect to my fundraising tech stack?
Despite all those connectors, there is not yet a single read-write major fundraising platform integration in any of these AIs. The closest it gets is the just-announced integration for Raiser’s Edge NXT to Microsoft 365. This integration actually plugs right into the Microsoft infrastructure layer but it does enable Copilot agents to answer questions straight from your RENXT data and you can build operations in Power BI to automate some database workflows. Blackbaud says nonprofits can request access to the integration but it’s not generally available yet.
Copilot also has a read-only connector for Salesforce but you need Power Platform’s Salesforce connector to write back.
The Agentforce Sales app is available inside both Gemini and ChatGPT, but it is built for the private sector and has limited capabilities for nonprofits.
Claude has a third-party connector for Salesforce through a middleware layer called Windsor.ai and a Raiser’s Edge connector that leverages the Blackbaud Sky API.
Before you implement an AI at your nonprofit
All of these options assume your organization’s data is not a hot mess and that you have done your due diligence figuring out your AI governance, access and data management policies, security posture, and privacy commitment.
We walk clients through a three-phase AI adoption methodology that comprises:
- Current state and goal definition
- Data infrastructure review
- AI recommendations and action plan
We use a maturity model to assess the organization and build an implementation plan that meets your org where it’s at and moves it toward its goals.
AI has huge potential to speed up nonprofits and help their teams make good decisions with greater confidence. Our goal is to help nonprofit leaders understand the risks, tradeoffs, and promise of AI for their organization so they can implement a solution that truly meets their needs today and into the future.
>> Need more help? Get in touch with our AI experts.
-
Director of Marketing
Lyndal has worked at the intersection of nonprofits and technology for most of her career, building strategic marketing programs and managing data-driven campaigns at the San Francisco AIDS Foundation, Nonprofit Technology Network, InfluxData, and others. She leads Heller’s marketing efforts and is excited to position Team Heller as the partner of choice for nonprofit and education advancement leaders. When not at her desk, Lyndal is usually on a hiking trail or listening to a podcast about star stuff.
View all posts
