By Dante Ricci and Lauren Gibbons Paul

Charities everywhere are feeling pinched. It’s a condition that is part economics and part donor fatigue. The U.S. economy, for example, is growing faster than expected but Europe and Japan remain flat. After a drop in U.S. charitable giving in 2022—only the fourth time in 40 years, according to a Giving USA report—donations in 2023 plunged further, by as much as 30%.

All of this makes new donor acquisition and donor retention, key imperatives, according to the Fundraising Effectiveness Project. According to NonProfitPRO, a resource for nonprofits, even recurrent donors feel fatigued and are keeping their wallets shut as they are hit with increased communications through different channels from individual charities and appeals from additional organizations. Given these trends, nonprofits can reasonably expect lower donation numbers, at least for the near future.

Ashutosh Nandeshwar, a consultant who advises nonprofits, sees limitless potential for artificial intelligence as a means to mitigate and reverse these trends. Nandeshwar, SVP of data science and analytics at CCS Fundraising, says nonprofits can use both traditional AI (including machine learning, natural language processing, and predictive analytics) and modern AI (including natural language generation and generative AI) to make more strategic use of limited resources.

Money–fundraising–is a core function of a nonprofit, like revenue generation in the private sector. However, these organizations face additional challenges beyond retaining generous donors and attracting new ones. Aligned with this need is the drive to strengthen the efficiency of their business processes and improve their mission performance. Leaders working to address these challenges are finding encouraging signs that AI tools, from machine learning algorithms to ChatGPT, can help.

“Especially with generative AI, not only can it be a lever for becoming more efficient, but [nonprofits] can also free up time to do the things they do best, which is being donor-centric, stewarding and cultivating those donors,” says Nandeshwar. “AI can help elevate philanthropic outcomes and, ultimately, better mission fulfillment.”

Artificial intelligence tools in a variety of forms can increase efficiency (save money) by quickly identifying potential new donor pools, analyzing past giving behavior, and zeroing in on the specific individuals and foundations who are most likely to donate. Nonprofits are also using AI for robotic process automation to eliminate costly and time-consuming manual processing of tasks, such as extracting data and filling in forms. Another important advantage: The ability to create personalized donor experiences that cut through cluttered inboxes and limited attention spans goes a long way toward perking up donors.

Charities may recognize AI’s potential, but we’re still in the early days of adoption, according to a survey by NonProfitHR, a talent advisory service. Its 2024 survey of 250 social sector organizations found respondents viewed generative AI as a useful tool for everything from taking meeting notes to writing job descriptions and drafting social media posts. Still, 45% of respondents said they did not yet use generative AI, with 14% planning to begin using it within the next 12 months.

For those getting started on AI projects, the opportunities are there and likely will grow with experience. To get the benefit of using AI tools, there are legal considerations to plan for, including privacy, regulatory issues, and insurance (see sidebar) But there is no question: AI can help nonprofits better manage their money, mission, and people.

Nonprofits have begun pursuing a range of projects in the past year-plus since ChatGPT and other widely available generative AI tools emerged in late 2022. Among the opportunities they have identified: AI helps nonprofits better understand their donors’ giving patterns.

Nonprofits are increasingly using generative AI to analyze donor data and predict giving patterns, resulting in more targeted and effective fundraising campaigns. For example, the American Cancer Society applied machine learning in a 2022 project to pinpoint which of its digital advertising campaigns generate the most revenue, as well as raise general awareness.

The program has generated impressive results, according to a press release from an advertising technology company working with the cancer charity. “Every bit of our campaign spend needs to be optimized for the best possible performance, so our key advertising goal was to reach the most probable donors, and then engage them in a way that would drive donations,” says Ben Devore, director of media strategy for the American Cancer Society. The program generated donation revenue that came in at 117% over its benchmark and a donor engagement rate of nearly 70% throughout the campaign. The click-through rate on interactive banners was 1.5%, 87.5% above the charity’s benchmark.

Opportunities to better understand and communicate with donors

Learn more about implementing AI

“8 Examples of Artificial Intelligence in Action.” Enterprise use of AI is growing fast. Here are eight examples with lessons for businesses of all kinds.

“A Playbook for Machine Learning Projects That Work.” Author Eric Siegel says most leaders miss the boat when it comes to deploying the biggest breakthrough technology to date.

“Prepping the Business for an AI World.” In this collection of SAP Insights articles, we reveal how AI is changing business and what it takes to realize its full potential.

AI-powered chatbots can streamline communications and increase donor engagement.

Generative AI can seamlessly create new content (such as e-mails, press releases, appeal letters, blogs, and social media posts) and foster creativity. The key is to keep a human involved for oversight of the tone and to ensure accuracy. This basic risk-reduction strategy is a good rule of thumb for all AI use.

AI chatbots can help guide potential donors through the donation process.

And they can help in designing and delivering personalized donor experiences. The tools make this quick and cost-effective to do.

Generative AI can fill in as your grant writing “co-pilot.”

Tools like ChatGPT are famously helpful as a starting point for communications of all types. But they are useful for higher octane work such as grant proposals, too. Services such as grantboost.io have emerged specifically geared to help nonprofits with grant writing.

Let generative AI help you cast a wider (foreign language) net.

The translation capabilities in ChatGPT and other tools make it easy and fast to engage with nonnative English speakers. Just be sure you understand the regulations when operating in a new jurisdiction (see below for more information).

Machine learning tools can help nonprofits unearth useful nuggets from the trove of data they possess.

This includes how their organization is meeting its mission and making a difference. Donors increasingly want to know more about where their donations go; insights driven from machine learning (ML) can provide the proof points they seek. This style of increased transparency should help decrease donor fatigue.

ML can point to receptive new audiences. St. Jude Children’s Research Hospital, which funds close to 80% of its US$1 billion operating budget through individual donors, was driven to find new ways to interact with existing donors and identify new ones. According to a story on the Future of Marketing site, St. Jude integrated ML into its emotionally engaging donor appeal videos and offered easy ways to donate in the moment and engage new, younger, and more diverse audiences, according to an article at Google’s Future of Marketing.

ML can boost the efficacy of a charity’s mission.

Perhaps the most exciting use of AI and ML is to help the nonprofit more effectively carry out its core mission.

Case in point: the Hebrew Immigrant Aid Society (HIAS), uses ML to pinpoint the best location for refugee resettlement while accounting for a refugee’s ability to speak the local language and find a job. A study found the tool could increase the refugee employment rate by as much as 40%.

"AI can help elevate philanthropic outcomes and, ultimately, better mission fulfillment."

–Ashutosh Nandeshwar, SVP of data science and analytics for CCS Fundraising

In another example, Breastcancer.org—a leading digital resource for people affected by breast cancer—uses ML to personalize the patient education journey, according to a NonProfitPRO article. The same article cites Polaris, a 150-person nonprofit that operates the U.S. National Human Trafficking Hotline, providing 24/7/365 support for survivors of human trafficking. With the number of calls from other agencies and survivors themselves increasing thousands of percentage points from 2008 to 2021, the nonprofit implemented a voice bot powered by ML to direct non-urgent general information requests (over 1,700 in the first six months) to informational resources, leaving staff free to engage with more urgent calls.

Gain insights into donor sentiment

Tools that incorporate natural language processing (NLP) and ML can sift through social media posts and reams of other structured and unstructured data to analyze how members of the public think and feel about relevant issues.

For example, the White Ribbon Alliance, a Washington, DC-based global women’s health advocacy group, used NLP to analyze findings in “What Women Want,” a survey of one million women from eight countries that asked about their reproductive healthcare priorities.

Responses from women in Mexico, Cameroon, Nigeria, India, Pakistan, Malawi, Uganda, Kenya, and Tanzania came back as unstructured data: text formats, using open-ended questions in a variety of languages, which made the survey difficult to analyze on a global scale. NLP made manageable the task of sorting through and understanding the responses to questions such as “What is your one request for quality reproductive and maternal healthcare services?” (Manual tabulations would have been unworkable with this number of respondents.) The White Ribbon Alliance shared the key concerns with case workers on the ground in respondent countries, and they were surprised by the depth of the data, which they were then able to use to help local constituencies, according to a blog post on Fast Data Science.

Legal considerations to nonprofits’ use of AI

Are you fired up to use generative AI or other tools to increase efficiency and donor engagement? Not so fast. First, there’s a host of legal issues you’ll need to consider. According to an article in the NonProfit Times:

Data privacy regulations that predate the debut of the newest AI tools are a top concern with AI use for nonprofits as they need to ensure compliance related to fundraising activities (for example, GDPR and the California Consumer Privacy Act), consent, safeguarding data, and providing individuals with rights regarding their personal information and restrictions on solicitation practices.

Intellectual property becomes an issue when AI systems create new works that need to be protected (such as software, content, and images). You will also need to take steps to ensure the AI system is not infringing on another intellectual property (IP) owner’s rights.

Discrimination can inadvertently be introduced by AI, either by perpetuating pre-existing biases into the system or by introducing new biases from the AI system itself. Either way, you’ll need to guard against unfair bias.

Tort liability can arise if the AI system emits inaccurate, negligent, or biased results that harm donors or prospects (or anyone else who is harmed by the content). Use human gatekeepers to ensure the accuracy and reliability of the AI system’s output.

Insurance is needed to reduce the legal risks identified above. Traditional directors and officers insurance is likely to be insufficient to cover these risks.

Rewards, but not without risk

Along with the opportunities offered by AI technologies come challenges, including legal considerations and regulatory risk, especially in regions with strong privacy laws like Europe’s General Data Protection Regulation (GDPR). The requirements to maintain data anonymity are growing, even in the United States, and that will affect your ability to leverage donor data.

Data quality is another challenge. Quality and availability of data are imperative. AI is only as good as the data it uses to learn and refine its outputs. Organizations often run into roadblocks with an inability to pull out data from old systems or difficulties embedding AI into legacy systems.

Equally important are the algorithms used to process and analyze the data. Plan to put a lot of consideration into the choice of algorithm.

You should also consider the possibility of unwittingly injecting bias into your use of AI and the unintended consequences that could result. It is important to fully understand the system’s limitations and to ensure AI algorithms are transparent and explainable. Ensure that the security capabilities provided by vendors provide appropriate encryption and access controls and that you can adhere to applicable data protection regulations.

You also need to make sure your organization has the right skill sets for the type of AI. Even if you are purchasing and using pre-built products, you should have some internal people with AI skill sets to be able to govern its use.

Managing these risks is prudent–but doing so should not close the door on the promise that AI tools have to help nonprofit organizations reverse recent trends.

“We have observed philanthropic trends for generations,” says Greg Hagin, principal and managing director for CCS Fundraising. “Over the last twenty years, donors down, dollars up. More recently, donors down, dollars down, factoring inflation. Whether a crisis or concern, we do believe these AI-enabled technologies can help address the challenge by strengthening connection and commitment with personalized messaging and tailored offering.”

Where to begin

For nonprofits interested in using AI to reap benefits relating to money, mission, and people, consider these starting points:

Start with small projects. Test the feasibility and benefits of the technology and then increase use as you become more comfortable. Jump-start the use of AI with pre-trained programs by investing in pre-built products to test out narrow use cases.

Collaborate cross-functionally. Include a cross-functional team in all the decisions and implementation to ensure the use of technology aligns with your mission and benefits the communities you serve. Don’t forget to engage with other nonprofit organizations to share knowledge, benchmark, and understand lessons learned.

Safeguard. Allow for learning and adjustments along the way, which will help minimize risks. Fully understand the systems’ limitations and ensure AI algorithms are transparent and explainable.