How I’ve been using AI chatbots in my Development Work
Photo Credit: Image by Rudy and Peter Skitterians from Pixabay
Amidst a lot of hype across all sectors about the power of generative AI, I’ve been slowly trying out different ways to put these new tools to use in my work. I’ve found a handful of ways AI chatbots have helped me be more efficient, informed, and effective as a fundraiser in my role with the Wisconsin Chamber Orchestra.
I’d like to share a few of those with you. I hope they spark your curiosity and desire to experiment with ways these tools might enhance your work, too.
And really, I’ve only gotten into the well-known, freely available generative AI chatbots found online, such as ChatGPT, Microsoft’s Copilot, and Google’s Gemini (I’ve used the first two). I won’t get into image generation—and I’m definitely not qualified to say much about more intense uses of AI.
I have three examples of little, specific ways I’ve used Copilot and ChatGPT that you, too, may find useful.
One-upping the Microsoft Office We Know and Love
In the past, when getting into some trickier data analysis or trying to build a clever dashboard in Excel, the best way to learn new skills (formula functions, macros, etc.) was to wade into the mystical world of user forums and 11-year-old YouTube videos posted by guys in their basements.
Now, the AI chatbots have those answers way faster and with way more specificity. I describe what I need in layperson’s terms, tell the bot the exact cells my data is in, and quickly iterate tweaks when v1 doesn’t give me what I expect.
It’s a similar case with grammar questions, including those that don’t seem to compute for Word’s spell check. Where I used to look at writer tips sites for examples of the grammar rule in question, I can feed the AI chatbots my exact text and ask for a detailed, short, or “really-dumb-it-down-for-me-please” explanation of what I got wrong.
A Quick Gut Check, with Sources Cited
These chatbots are fast and have huge swaths of the internet behind them. I’ve found Copilot helpful in heat-of-the-moment conversations where my professional gut instinct throws a flag.
Picture this scenario: In a gala fundraiser committee meeting (via Zoom), an enthusiastic committee member promotes having as many silent auction items as possible: 200 as a minimum. That doesn’t ring true to me – isn’t that too many? The rest of the committee nods along – am I the crazy one?
In real time, I can ask Copilot this: “hey, for big fundraiser gala (500 guests), what would be considered a ‘standard’ number of silent auction items?” In this case, Copilot offered me info from redappleauctions.com (a site I hadn’t heard of before) recommending a formula of one auction item for every two “bidding units” – 125 for my hypothetical event.
Will I stick with this number? Maybe, maybe not. But at least I have quick external validation for why “200 minimum” raised a flag and, while the conversation is fresh, I can steer my dedicated volunteers to consider a lower goal number.
My Donor’s Expertise, In My Novice Vocabulary
In my job, I feel lucky to get to meet many donors with serious academic backgrounds who may have had 40-year careers as academics at UW or Edgewood. Like we all do, I try to be prepared before meeting with them, but getting a general idea of their research area can be difficult when they’re a career-long organic chemist and I didn’t exactly ace senior year Chemistry.
So, after Googling my donor prospect and seeing they published a work called “Synthesis of Ether-Protected Alcohols Using Hypervalent Sulfur Reagents,” I plug the article title and abstract into Copilot and ask it to give me the gist as if I’m a precocious eighth grader. It won’t retroactively improve my science ACT score, but now I can do one step better than just nodding and smiling when the donor talks about their research passion.
Watch Out, Though
The responses I sometimes get from Copilot and ChatGPT can sometimes be quite off-putting in how “people-pleasing” they are. The responses are often over-enthusiastic, sprinkling exclamation points freely and describing my questions as “fabulous.” If I’m not consciously avoiding it, it can be easy to use these chatbots as an echo chamber instead of an outside party.
Extending my “number of silent auction items” example above, when I instead quizzed Copilot for “which of us is right, the volunteer with 200 items or the professional with 50,” it took my side right away and spent two sentences talking about how 50 makes great sense (with no online source cited). Instead, I’ve gotten the most valuable answers when I’m value-neutral in my question, not seeding my prompt with any particular numbers or pre-assigned idea of high or low.
And, of course, it has to be said that most of my experience with writing fundraising copy via these chatbots has been poor. Don’t expect anything but hyperbolic, vapid shadows of real, emotionally grounded fundraising writing unless you give the bot a major amount of prep information (i.e. heavy outlines with the copy half-written) and prepare yourself for major editing afterward. I’m sure they’ll get better – and I’m keeping an eye out for fundraising-specific tools – but they’re not there yet.
Lately, Copilot and ChatGPT have become go-to tools in my toolbox; ones I probably use at work every day at this point. If you haven’t taken that little leap of faith yet, now’s probably the time to try one out for a small, discrete question like those I posed above. I think you’ll surprise yourself with the uses you find for it.