Demystifying AI for Nonprofits: How Does it Work?

Leadership

Dec 22

We see it everywhere, everyone’s talking about it—but what does AI actually do for nonprofits?

Maybe your organization has been excited about the AI buzz but is unsure of how exactly this technology might fit into your work (or the potential risks this new technology presents). You’re not alone.

A baseline understanding of artificial intelligence tech will be essential for success in all kinds of industries and sectors going forward, nonprofits included. AI tools can help your organization unlock new levels of growth and efficiency, but only if they’re being used properly.

If your nonprofit is ready to make a change and invest in new tools, what should you know?

This crash course will orient you to the basics and walk through some of the key risks and additional considerations that AI raises for nonprofits.

What is artificial intelligence?

In the DonorSearch guide to nonprofit AI, we define artificial intelligence as “the ability of a machine to ‘think’ like a human and perform tasks like recognizing patterns, processing information, drawing conclusions, and making recommendations.”

For most organizations and businesses that want to use AI, there are two general types to be aware of, differentiated by their purposes and the forms of their outputs:

Generative AI. These are tools designed to generate written or visual content as responses to prompts. They use large language models (LLMs) to interpret the nuances of human language and improve their understanding over time.
Predictive AI. These tools work with more limited sets of data to generate statistical predictions about future actions or results. Most of the platforms designed specifically for nonprofits fall into this category.

Simpler forms of AI have been with us for a long time, like basic help chatbots on websites and predictive recommendation algorithms used by businesses like Amazon and Netflix. But advances in machine learning, the ability of AI systems to actively adapt their “thinking” processes based on new data, have driven the recent leap forward in AI technology.

Generative tools like ChatGPT have stolen the show in the past couple of years, and they represent a huge step forward with their ability to interpret language. Predictive AI tools are also becoming increasingly effective at uncovering complex patterns and continuously improving their suggestions over time.

How does it work?

Artificial intelligence tools use complex processes to produce predictions and responses, but to boil it down into a few very general steps, we can think of it this way:

Data is provided to the AI system. For generative AI, this is often a very large set of publicly available data and content from the internet. For predictive AI tools, the data comes from your own database and/or a third-party collection of curated data.
The AI analyzes the data. This occurs on an ongoing basis or as prompted when new predictions are requested.
The AI identifies patterns and makes correlations. While analyzing the dataset, the technology finds deep patterns and relationships between data points over time. It develops models based on its understanding of these relationships.
The AI produces outputs. Using the models that it’s developed, generative AI will produce responses. Predictive AI will produce specific predictions by projecting its predictive models into the future.
The system adapts over time. Through the process of machine learning, the AI tools can continue improving and refining their models as they’re provided with fresh data over time.

Generative AI tools are often standalone applications that don’t require additional inputs from your organization. To use predictive AI software, though, you’ll need to provide the system with your nonprofit’s own in-house records of donor information and engagement histories. In many cases, you’ll also augment your data with external datasets curated and maintained for a specific purpose, like a prospect research or wealth screening database, or a smaller data append package.

What are AI’s use cases for nonprofits?

So, how exactly are nonprofits currently using AI technology to improve their work and drive results? These are some of the most prominent use cases, separated into those that help internally and those that are visible to donors and external stakeholders:

Internal Use Cases for AI

Generating predictions. As mentioned above, AI can analyze your data to give you specific predictions about future donor actions, like who’s likely to give within the next year or generate the greatest lifetime value for your organization.
Screening and segmentation. AI can also screen your database to help you refine your segmentation and qualification strategies, like by identifying new prospective major donors who weren’t previously on your radar.
Maintaining data and automating routine tasks. By combing through your data, AI can help you maintain its hygiene and usefulness over time. Use it to generate data visualizations, eliminate duplicate or outdated information, and automatically update donor profiles.
Detecting fraud. With an understanding of what typical transactions and interactions look like for your organization, AI can flag fraudulent activity that needs to be investigated.

Donor-Facing Use Cases for AI

Online chatbots. You’ve probably already encountered these help bots on other websites. Many nonprofits already use them to help direct visitors around their websites and encourage donations.
Generating messages. Use tools like ChatGPT to develop templated versions of emails, text messages, and social media posts that you can adjust and deploy to quickly get the word out about your mission.
Personalizing appeals. Tools that integrate with your database can go a step further and truly personalize your appeals to donors based on their previous engagement histories and donation amounts.

Organizations of all sizes are today tapping into use cases like these to implement smarter, more agile fundraising practices. AI makes it possible to save time and strengthen your strategies on multiple fronts, from qualifying and prioritizing prospects to generating quick, reliable outreach lists to reducing the time needed for logistical tasks.

What new considerations and risks does AI bring?

Of course, every new technology brings risks and concerns. The organizations that use them, accepted best practices, and regulatory norms must all evolve to meet them over time. AI is no different.

First, artificial intelligence brings some ethical concerns, revolving primarily around the data that it needs in order to work. How this data is collected and how it’s ultimately used have raised numerous concerns about privacy, inadvertent bias, and cybersecurity.

Additionally, AI can produce outright incorrect or inappropriate predictions or responses (here’s one particularly bad example). Never simply copy and paste an AI response to be used in public-facing communications without first reviewing it and making changes to align it with the context and your organization’s voice. Carefully monitor any chatbots on your website and establish clear protocols for when it should direct visitors to speak with a human staff member.

What are the steps you can take to mitigate AI risk?

There are plenty of steps that you can take as an organization to reduce the potential risks of using artificial intelligence to guide decisions and engage with donors. Here are our top recommended practices that should be integrated into your overall risk management strategy:

Carefully vet your AI and data vendors to ensure they’re using ethically sourced data and have set up robust security protocols.
Revisit and expand your own security practices to protect sensitive donor and constituent data.
Educate your whole organization on the basics of ethical AI usage, and provide thorough training to anyone who will directly use your AI tools.
Stay on top of regulatory developments and emerging best practices in the AI space.
Center human involvement in all AI workflows—the expertise and judgment of your own team should always be involved in final decision-making.

For a deeper dive into these steps and the Responsible AI Framework that’s becoming an industry standard, explore our guide to responsible AI usage for nonprofits.

With a strong foundational understanding of artificial intelligence and how nonprofits might use it, you can set up your organization for success in this new era of fundraising. This technology can completely improve how you identify, understand, and engage with your donors, ultimately saving you more time that can be put towards building stronger relationships and delivering on-the-ground impact.

Sarah Tedesco, Executive Vice President of DonorSearch

Sarah Tedesco is the Executive Vice President of DonorSearch, a prospect research and wealth screening company that focuses on proven philanthropy. Sarah is responsible for managing the production and customer support department concerning client contract fulfillment, increasing retention rate and customer satisfaction. She collaborates with other team members on a variety of issues including sales, marketing and product development ideas.