What the Future Might Hold for A.I. and Fundraising

Ashutosh Nandeshwar can see a not-too-distant future when sophisticated chatbots answer potential supporters’ questions online and connect them with the fundraisers most knowledgeable about the causes they care about. Donation pages on university websites could serve up the five or six funds a donor is most likely to support, based on that person’s interests, instead of the hundreds that donors currently confront.

It’s not that far-fetched a vision.

Last month, Ruffalo Noel Levitz, a higher-education software and consulting company, released A.I.-powered chatbots that universities can use for fundraising and enrollment.

“It really allows for an institution to continue to communicate with a donor, with an alum, with a friend, with a parent 24/7 regardless of whether someone at that the institution is there or not,” says Josh Robertson, the company’s senior vice president for product strategy.

Using artificial intelligence to predict which issues will spark a donor’s passion could also help nonprofits make giving easier, says Nandeshwar, an assistant vice president at the University of Southern California and co-author of Data Science for Fundraising.

QuadWrangle, a company that serves higher education and private schools, offers an online system that uses machine learning to understand alumni, drawing on data from the institution’s database, their behavior on social media, and more. It then uses those insights to select what news and stories to show donors when they visit the university’s website or to include in the emails the institution sends. The choices donors make as they interact with that content then helps the system learn more about their interests and improve the personalization.

Each week, universities and other large nonprofits create pieces of content by the dozens, if not hundreds, says Nick Zeckets, co-founder and CEO of QuadWrangle.

“With all that volume, we only have to do a great job two or three times a week,” he says. The goal isn’t to replace someone’s Google news feed. “We’re just trying to create a great level of resonance between an institution and one of their constituents regularly enough that you always feel relevant as an institution.”

The system serves up tailored giving options when donors visit the university’s website, and it also embeds very specific donation appeals into stories that take people to a specific giving page.

A group will get more click-throughs — and ultimately more gifts — if the story itself doesn’t include a donation plea, Zeckets says. “People don’t like being sold to; they like discovery.”

Pulling Data Together

QuadWrangle and other A.I.-driven systems designed for colleges and other large nonprofits, like Uprising Technology and NewSci, bring together and organize a group’s data from different locations — which nonprofits have long struggled to do, making it difficult to get a full picture of donors’ interaction with the group.

Imagine that a volunteer managed an event and checked people in using a spreadsheet, Zeckets says. “Did they use the unique user ID from the school’s core database? Did they write that down, too? Of course, they didn’t.”

What normally happens in a case like this, he says, is that fundraisers use the spreadsheet to send out thank-you notes, but the information about who attended seldom makes it into the main database.

Donors know all the ways that they’re involved with a nonprofit, and they get frustrated when the organization doesn’t seem to know the full extent of what they do, says David Lawson, co-founder of NewSci.

A big part of the problem, he says, is that nonprofits have struggled to work with both structured data — think concrete fields in a database, like gift amounts — and unstructured data — free-form information like the notes a fundraiser entered into the database after visiting a potential donor. Artificial intelligence is uniquely able to do that, he says.

“This technology is built to pull that together and get these organizations on the same page as the donor.”