AI can do a lot of things these days. It can write a sonnet or a song, create an abstract picture or a photo-realistic rendering, and save us valuable hours each day while making us fear for the future of the human workforce.* But what can AI do when it comes to data—and what are its limitations? We’ve got your go-to guide on AI platforms and nonprofit data right here.

*It will all be okay. Really! According to the World Economic Forum, AI will lead to the elimination of 85 million jobs … but add 97 million new ones. That’s a net gain!

First, let’s ask Chat GPT: What can you do when it comes to data?

Well, we asked our friendly AI colleague the question above, and it was a bust … it just went on and on with vague lists of what AI can do. So, we got more specific and asked, “Can you show AI your fundraising reports and have it analyze or decipher them?” Then, we had it shorten its response by a solid 75%. (Chat GPT is verbose; what can we say?) Here’s its final answer:

Yes, AI can analyze fundraising reports and summarize them by using natural language processing (NLP) to extract key information, creating visualizations to highlight trends, using predictive analytics to forecast future behavior, and generating summary reports in natural language. This enables nonprofits to make data-driven decisions to improve their fundraising efforts and achieve their mission.

Cool, cool, good to know! Let’s get into the details.

Chat GPT’s capabilities and limitations for data analysis

It sounds like Chat GPT and its brethren can really help out when it comes to your fundraising data. But what, exactly, can it do—and where will you need to turn to a human analysis? Here’s an overview.

• AI platforms can process and analyze large amounts of data and identify patterns and trends. When asked, they can also provide insights and recommendations and apply machine learning to make predictions.
• AI is quite helpful when it comes to cleaning up data. For example, it can check for consistent formatting or highlight missing values. But it cannot account for flawed data. So, if you feed it incomplete or inaccurate data, the analysis will be flawed.
• While AI can certainly display some very human biases, it typically won’t make any recommendations or decisions that are subjective. Therefore, all of its suggestions will be based on the data—not on moral or ethical considerations.
• AI has no context for your data beyond what you give it. And because AI platforms only incorporate text data until 2021, you need to train them on anything newer. To get the most relevant insights from your data, share all the latest and greatest updates on your nonprofit and industry first. For example, you can “tell” Chat GPT about your organization and industry by pasting in updated content from your website and sharing the newest, most relevant articles.
• You can only upload text to Chat GPT, Google Bard, Jasper, and their ilk. So, that means you can’t directly generate graphs or charts within the interface (though it will tell you otherwise!).
• To analyze data in Chat GPT, it has to be in a text format. And happily, that’s pretty easy! Just paste those relevant rows directly from your spreadsheet or convert them to a CSV (comma-separated values) file.

A play-by-play of AI data analysis for nonprofits

To save you a lot of back and forth with your AI consultant of choice, we spent some time data-and-chilling with Chat GPT. Below, check out the steps we took to get the analysis we wanted.

1. Clarify
2. Start high-level
3. Ask more questions
4. Request recommendations
5. Wrap it up
6. Convert the data
7. Support your findings

Step 1: Clarify

If you drop in a data set (say, containing basic fundraising donor data, such as demographics and giving history) with no context and ask for an analysis, it will ask for further clarification or suggest ways to narrow your request.

Step 2: Start high-level

If you ask for general insights on the data, Chat GPT can give you a list of key takeaways. For example, it can tell you:

• Whether you’re missing information for any key areas.
• The distribution for your donors in any categories (gender, marital status, etc.)
• General insights about giving. Chat GPT told us, “Some donors have given consistently over time, while others have not given at all or have given sporadically. This information can be useful for identifying trends and predicting future giving behavior.” *’scuse us while we eyeroll*

Step 3: Ask more questions

To get deeper insights, ask more probing questions. You can ask it to do some calculations, such as calculating donor retention rate based on giving history (it will even tell you how it calculated it, which is good because you always want to check that math!). Or, you can ask it to identify key trends from your data, and it will discuss how your donor retention rate has changed and what that means.

Step 4: Request actionable recommendations

Be sure to consistently refer back to the data (e.g., start your queries with “based on the data above”) or it may revert to more generalized suggestions.

Step 5: Wrap it up

From there, you can ask the AI to rephrase the information for your upcoming board meeting or for your annual report. Ask for a one-paragraph summary, a numbered list, or a detailed explanation. Be sure to specify your audience to inform the tone.

Step 6: Convert the data

If you want to visualize any of your data, you can provide it in text format and Chat GPT will direct you to appropriate tools and explain how to convert your data into the desired format. (Or you can just let Funraise’s fundraising intelligence feature do all of that for you automatically, on brand, and with one click.)

Step 7: Support your findings

If you need external data sets for context or to help you up your fundraising game, Chat GPT can point you toward organizations and/or websites that are likely to have just the data you want.

Chat GPT’s suggestions for how you can help it rock data analysis

Additionally, we wanted to give Chat GPT the opportunity to tell us, in its own words, how we can help it succeed when it comes to data. Here’s what it had to say.

Provide clear and accurate data: The more accurate and complete the data you provide, the better my analysis will be. Please make sure to double-check your data and correct any errors before sharing it with me.

Give me context: It's important to provide context around your data. Let me know what the data represents and what it means for your organization. This will help me understand the data better and provide more relevant insights.

Ask specific questions: The more specific the questions you ask me, the better my answers will be. Don't be afraid to ask detailed questions about your data and what it means for your organization.

Provide feedback: Let me know if my analysis was helpful or if there are areas where I can improve. Your feedback will help me better understand your needs and provide more useful insights in the future.

Keep me updated: Keep me updated on any changes to your organization or data. This will help me provide more relevant and up-to-date insights.

Types of fundraising data analysis Chat GPT can do

As you can see, AI can make data analysis a heck of a lot easier, even with some significant limitations. What, exactly, can it do for your fundraising data? Here are some fundraising analyses where AI can be a great wingman.

1. Revenue and expense data
2. Donor data
3. Fundraising channel analysis
4. Campaign data
5. Fundraising trends

Revenue and expense

Wondering how your revenue’s balancing out? AI tools can help you understand your financial health and identify areas where you may need to cut costs or lean in a little more.

Donors

AI platforms can look at your donor data, analyze their characteristics, and help you understand who your most loyal donors are. In addition, they can help identify opportunities to engage new donors.

Fundraising channels

AI can review your different fundraising channels to help you understand which ones are most effective at generating revenue and engaging donors.

Campaigns

Your AI interface of choice can review the performance of specific fundraising campaigns to help you understand which campaigns were most successful. Then, you can ask it to suggest ways to improve campaign performance going forward.

Trends

If you want to understand how your organization is changing and growing, ask your AI helper to analyze your fundraising data over time, such as year-over-year revenue growth, changes in donor demographics, or shifts in giving patterns.

What about Microsoft Copilot and other new AI tools?

In these nascent days of AI, every hour brings new technologies and updates. Most recently, Microsoft announced their Copilot technology, which will integrate with the Microsoft 365 apps, including Excel. With Copilot in Excel, Microsoft explains, “you can analyze trends and create professional-looking data visualizations in seconds.” Google, too, has released Simple ML for Sheets, a machine learning integration, which is now in beta. Stay tuned for additional AI integrations and updates! They’re sure to make our busy nonprofit lives less busy in many ways … and busier in others. It’s all part of the journey, folks!

AI and data for nonprofits: Key takeaways

• When it comes to analyzing data, AI can do a lot, but it has limitations.
• AI can analyze fundraising reports, clean up data, identify patterns, and provide insights and recommendations based on that data.
• To analyze data with AI, the data has to be in text format and the data needs to be accurate. Inaccurate data will lead to flawed analysis.
• You can help AI provide better analysis by providing it clear and accurate data and as much context as possible. You can also ask it specific questions and provide feedback.