I have written quite a few articles about chatbots and no/low code AI in the past year but have seen nothing quite as promising at this moment in time than ChatGPT. After spending the past few days playing around with the platform which was launched this week by the OpenAI group and is already over 1 million users to date I wanted to share some of the potential benefits of it for our nonprofit community.
ChatGPT is a natural language processor which in simple terms could be seen as a text generator based on user prompts. Think of a combination of Google search and those AI image generators (that were all the rage last month) but instead of ranking links, indexed answers, and templates to the question posed, it provides a nuanced answer beyond the binary nature of inputs and outputs that we have been used to the past decade. It is capable of understanding natural human language and generating thoughtful human-like prose after a prompt.
In short, you can just type in a command and the AI will create something based on those parameters. I’ve already seen a few posts on social media saying that they got a fundraising plan and a donor thank you email from queries they made on the platform.
Exciting stuff right? Well not to the levels we should be thinking of, because that’s just scratching the surface and our sector would be wise to dig deeper into the platform to discover its real potential.
Yes, it has the potential to automate some of the copywriting that consume much of our time, which can include cases for support, grant writing and the day to day email correspondence with our members, and yes, that should be applauded, but with a bit of trial and error I have stumbled on to some more substantive applications.
The first thing I realized is that if you just use it as a text generator you will just get text as a response. This is no fun and if you use ChatGPT to draw down data to help build out a case for support then it will tell you that it is unable to browse the internet. But I found it is quick to eat up new data including direct html links to search terms and .csv docs which opened quite a few doors (and provided me with the code to unlock them).
Prospect research: you can basically scrape LinkedIn for the name, companies, roles and URL profiles of folks that went to a particular college or work at a company that has a corporate matching program, layering it with other keywords.
Other data sets: pull countless lists including search options like grantmaking foundations from 990’s using name, IRS taxonomy codes and location in relation to you in miles.
Dashboards: a great thing about ChatGPT is that you can also utilize other apps within the commands. This includes dashboards and data visualization which you know can tell your stories of impact better through quality and aesthetic tweaks. It can also build good UX dashboards to build out 3 years prospect pipelines etc.
These are just a couple of options that have been explored. There are many, many more and that’s the beauty of the platform and why it’s best to leave this post here – as a primer, not as a blueprint.
Is the revolution in AI now here? Not quite, but it’s fun to have a true usable insight into what it might do for our sector and we would be wise to engage with it now, not only to learn what’s possible, not only to potentially make our jobs easier today but to ideate and provide feedback for its improvement. The better it gets, the better the outcomes. After all it’s a platform which is optimized for dialogue and we would be remiss to be a silent bystander while tech continues to evolve at speed.
Future Philanthropy Blog 