Artificial intelligence has a hand in just about every industry possible, including the nonprofit sector. Though some may not realize it, nonprofit organizations account for nearly $400B of revenue annually worldwide. But some nonprofits still struggle to make ends meet and expand their workforces, causing them to fold and lack sufficient funding from donors.
David Woodruff, VP of Resource Development at Massachusetts Institute of Technology, and Rich Palmer, Co-Founder of Gravyty, spoke about how to boost efficiency at nonprofits using predictive metrics and machine learning at ODSC East’s Accelerate AI Business Summit Wednesday.
Automation and artificial intelligence are two areas many feel threatened by for reasons ranging from job security to societal decline. These worries extend to the nonprofit workforce as well, especially considering nearly 12.3 million Americans are employed by nonprofit organizations, according to a report from Johns Hopkins University. But Woodruff and Palmer don’t see AI as a deterrent for growth—in fact, they find it essential.
“The nonprofit sector, despite its best efforts, is usually about ten to twenty years behind in adopting new technologies,” Palmer said.
For nonprofits, the revenue that comes in goes right back into funding the company’s essential needs. Theoretically, more revenue would also mean expansion within the company, and the potential for creating jobs. With Gravyty, Palmer hopes to use machine learning technology to help nonprofit workers perform more efficiently and generate revenue, rather than cut them out and leave them hanging to dry.
Most often, 95 percent of a nonprofit’s revenue comes from only five percent of its donors, according to Woodruff. This makes it easy to track metrics and create close relationships with these top donors. The problem is, uncovering insights on the middle 80% of donors is much harder with such a wide variety of cases. To remedy this, Palmer and his team at Gravyty thought through the problem and arrived at the idea any great data science team would—they built a machine learning model.
Palmer’s model uses an array of open source tools for machine learning, including spaCy, Sci-kit Learn, TensorFlow, and the Natural Language Toolkit to provide insights that boost strategic thinking. By analyzing interactions and relationships with donors from across the board, workers are able to better plan how and when to reach donors for best results. Using these predictive models, nonprofits at the College of Charleston and the University of Delaware have seen noticeable changes in efficiency. In fact, Gravyty estimates as much as a 160% increase in workforce expansion.
The most noteworthy takeaway of the introduction of AI in nonprofits is that minimal change was necessary to boost revenue. In fact, the amount of required work necessary was reduced significantly, allowing employees to direct their attention to more pressing matters. With an uptick in revenue, the doors are open for the option of creating jobs that didn’t exist previously. So are robots going to take over industries in the near future? I wouldn’t go that far, but one thing is for sure—they have the potential to make life a whole lot easier.