Sophie W. Penney, PhD
Artificial Intelligence and Fundraising: Is a Bot Coming to Take My Job?
Sing along with me: AI, AI, O!
Oh my, you say. Why sing when a robot might be coming to take my job some day?
Artificial Intelligence, or AI, seems to be in the news almost everywhere today. Many of those stories suggest that robots are coming to take the jobs of many and that it might even eliminate whole professions, much as was the case during the Industrial Era.
Speaking of the Industrial Era, I hail from Western Pennsylvania. It experienced one of the greatest economic booms and busts as the Industrial Era exploded in the early twentieth century, then fizzled in the 1970s. The city of Pittsburgh, the largest city in Western Pennsylvania, is experiencing a new boom thanks to high technology. This is in part because the city is home to Carnegie Mellon University, a bastion of AI technology research and production. As a result of my connection to the region, I found myself listening to a few online offerings about AI which have been produced by CMU faculty. Hearing that 25 percent of people might one day be out work because of this new technology was unsettling, to say the least.
Being a fundraiser, I asked myself, seriously, could a bot (or even a chatbot) take my job one day? In its review of 2018 charitable giving, Blackbaud noted that relationship-building continues to be the key to success (Charitable Giving Report: How Fundraising Performed in 2018, Blackbaud). So, you and I might ask if a machine could possibly do what we do and do it as well.
Before you say to yourself, “my job is safe,” ask yourself this question: Just because a bot might not be coming to take my job, should I ignore AI altogether?
The reality is that running from AI is like placing our heads in the proverbial sand. AI is here and is most likely not just an unwelcome houseguest that will leave sometime soon. No less than The Chronicle of Philanthropy made AI the focus of its March 5, 2019 special report titled, “A.I. and Fundraising: the Future is Here” (more about that in my next installment).
Still not convinced? Consider this: donors are increasingly using online and mobile devices to make gifts. While online giving was only 8.5 percent of all giving in 2018, it is consistently on the increase. Blackbaud’s report indicates that of those gifts, 25 percent were made using a mobile device. Even if your donors continue to give using a reply card, mailing envelope, and a stamp, AI can help your organization more efficiently and effectively interface with postage stamp and mobile generation donors (the latter of which will hopefully be your future major and planned giving donors). Finally, like it or not, AI is already being employed in fundraising – and with success. So, rather than run and hide, I invite you to do what I try do when exposed to something new: learn about it.
My first exposure to the use of AI in fundraising came as a result of attending a co-op meeting for public television fundraisers hosted by Tim O’Leary from Next Generation Fundraising, a consulting firm located near Philadelphia. One of the most fascinating presentations was delivered by Adam Treiser from Arjuna Solutions, a company that applies AI to help nonprofits raise more money. Adam, the founder, has a background in law and business and founded this company to help nonprofits more efficiently and effectively raise funds. Given that the mission of my firm, i5 Fundraising, is to help nonprofits more efficiently and effectively raise more money to change and save more lives, I sat up and took notice.
More recently I found myself on the receiving end of emails from another AI company named Gravyty. The emails shared statistics about significant increases in fundraiser productivity, the purported result of use of AI by a particular college. I subsequently learned that Gravyty has been featured in the Huffington Post which indicated that the company is “transforming fundraising.” Having led teams of fundraisers, I was intrigued by Gravyty’s work, yet at the same time wondered how the statistics were derived and the claims justified. That led me to contact staff at the company in an effort to learn more.
However, before I speak further about these companies and how AI is being used in fundraising, it seems important to be sure that you and I are on the same page with understanding what AI is (and what it is not). To learn more about AI, I enrolled in AI For Everyone, taught by one of the co-founders of online course content provider Coursea. In the introduction, Andrew Ng explains that there are two types of AI:
- Artificial narrow intelligence, or ANI (think Alexa or a self-driving car), and
- Artificial general intelligence, or AGI (in short, something a human can do, such as meeting with and discussing a gift with a donor).
According to the instructor, strides have been made with regard to use of ANI, but there is a very long road to travel before the robots take over for us (AGI).
I’ll provide an example. While watching a 60 Minutes segment several months ago, I heard an AI expert speaking about how AI is being used to teach children in rural China. He described many benefits but when it came to things such as reading a child’s facial expressions or understanding their motivations — all things that we as fundraisers do as we interface with donors — AI fell flat. An article shared with me by a former student speaks to this topic as well. A researcher set out to purposely fool AI by disguising specific objects as other types of objects, for example, overlaying a photo of what looked like a white golf ball on a white tea pot. AI could not distinguish the overall shape of the teapot and instead identified the item as a golf ball.
Needless to say, you as a fundraiser want very much to identify Mary Martin as the sixty-five-year-old longtime loyal donor that she has been. When you sit down with Mary you want to learn about her life and why she supports your organization. You want to read her facial expressions and hear the tone of her voice as you have this conversation; these types of interactions are where AI is far from ready to perform.
But what if before you met with Mary, you found yourself with a goal of upgrading a group of annual giving donors of which Mary is a member? If you have such capability you might run a query in your donor database so that you could segment the mailing (meaning send different messages to different groups of donors, e.g., millennials or boomers). However, how do you know if your segment is correct? How do you know what ask amount, if any, to include in the letter that is sent to Mary, let alone to everyone else who might be in the same segment. Your answer might be that you do it the way that you always have, which is to set a specific ask string (a list of possible gift sizes) and hope for the best.
However, you might wonder to yourself whether the mailing would be even more effective, i.e., raise more money if Mary’s specific giving history could be used to set the amount of the request. “Impossible,” you say. “How could I figure this out? Even if I could do this for Mary, I have hundreds of donors in my database. I could not possibly determine a specific ask amount for each of them.” Enter AI, which is being used for precisely this type of predictive modeling and development of direct-mail-appeal asks.
As I learned from Adam at Arjuna, AI can enable an organization that has sufficient and sufficiently organized data to determine an ask amount. Not only that, the data can also be employed to determine when it might be best to send a mailing to Mary. “Wait a minute,” you say. “Isn’t year-end still the best time to ask?” Blackbaud’s Charitable Giving Report, which I referenced earlier gives evidence that giving is becoming a year-round activity. June, which is the end of a fiscal year for many nonprofits, is now second to December as a significant giving period. To my great surprise, the report further indicated that July and August were peak months for mobile giving, with mobile gifts representing 27 percent of online gifts given in those months! What’s more, giving is spreading across all months of the year somewhat more evenly than many might suspect.
What does this tell us? That donors might not all be as motivated by year-end appeals as we might like to think. This point reminds me of a conversation I had with a donor some years ago, who was giving to a nonprofit that I served. We discussed the timing of the “annual” appeal. While my instinct and experience were telling me to mail a calendar-year-end appeal right after Thanksgiving, to my surprise he told me that he makes all of his giving decisions by October. I’ll provide another example: I have in the past made a memorial gift to a nonprofit in June of every year in memory of one of my sisters, whose birthday was in that month. The contribution amount has increased by one dollar per year because I made a gift equal to what my sister’s age would have been that year. AI has the potential to identify the right time to solicit me, the donor mentioned above, and others who contribute at varying times of the year.
Okay, you say, this sounds interesting, but does AI positively impact the fundraising bottom line, figures like donor retention, and total dollars contributed? Yes, and yes. Testing that has been done with real-world nonprofits shows that retention and gift amount have both increased as a result of use of AI. The overall result is an increase in donor lifetime value (if you are not familiar with lifetime value, read Adrian Sargeant and Elaine Jay’s book, Building Donor Loyalty: The Fundraiser’s Guide to Increasing Lifetime Value. What’s more, lifetime value did not increase over just one year; it continued to increase in subsequent years. You should also read The Donor Lifecycle Map: A Model for Fundraising Success, by Deborah Kaplan Polivy, to see how to retool your development efforts according to the donor’s lifetime giving cycle.
What are the implications for you and your nonprofit? The ability to capture and manage data, particularly data about the preferences of your donors, is critical (see this The Agitator article, “Don’t Talk to Me When I’m Not Listening”). From my vantage point as a consultant, many of the nonprofits I see losing ground with fundraising have antiquated databases, if they have one at all beyond an Excel spreadsheet. To be successful with annual giving and other forms of giving going forward, even if you do not employ AI, it is going to be critical to be able to gather and retain key data about your donors, why they support you, and when (including how often) they want to support you. Organizations that want to hold their ground, or to raise even more funds, will use this information to tailor appeals to donors. AI can smooth this path, particularly for small nonprofits that have lean staff structures.
Intrigued? In my next installment, I’ll talk about how AI is being used in major and planned giving fundraising. You will learn about how AI and tools such as a chatbot are being employed by users and the impact they are having.
We'll inform you about just-published articles, our upcoming books, professional Summits, live interviews, webinars, and more
Request Permission to Reprint This Article
Copyright © 1992-2020 CharityChannel LLC.