All right, I’m going to share my experience here, though this may be an unusual application of AI. This is long, sorry. But it goes unexpected places.
I started experimenting last week with having GPT give me strategic feedback on written drafts of grant applications. I put a lot of time into it - this is not time-saving in any respect. For context, I have fifteen years of experience writing grants and consider myself a serious writer. While GPT is often more concise than me, it is not a better writer. But what it can do is point out places in the Notice of Funding Opportunity that don’t align with my proposed project, and make suggestions for integrating funding priorities from the NOFO in specific places in my narrative. In a pinch, such as yesterday, it helped me cut 250 words from a narrative.
So based on that limited success, I asked it to create a self-guided course for me to learn how to integrate AI into my grants management workflow. By this point, it knows everything I’ve told it about the nature of my job, how long I’ve been working there. I asked it to ensure the coursework was ADHD friendly, and I provided it with an overview of my cognitive strengths and weaknesses, and this is where it took a turn. Suddenly the project focus veered from improving AI skills to addressing my specific executive functioning challenges within the context of my specific job.
Now I’ve read books broadly about how to deal with ADHD, and I’ve read books broadly about how to be a grants manager, and I even had an actual ADHD coaching group, but the level of specificity I got into with GPT was unrivaled. After some back and forth, GPT said, “I notice you said you miss things frequently. But elsewhere you say you’ve never missed a deadline. It seems like the reason you’ve been so successful is because of constant hypervigilance and all the tasks you have to keep in your head - You’re not forgetting things, you’re remembering everything all the time. This is probably why you feel so overwhelmed. So let’s create a strategy to reduce the burden on your working memory so you only have to hold one thing in your head at a time.”
I’m not gonna lie, I almost burst into tears. I’ve been well aware of the problem for years, and have tried countless things to solve it, including using Trello. So I told it about Trello and it advised me to stick with Trello, but with some tweaks. We created a category called the Parking Lot where I get to put everything likely to distract me, including creating new workflow systems. We talked about what, on Trello, I found overwhelming. And we were able to collaboratively hone in on a key insight: when I see the name of the project, I am immediately holding all of the project in my head at one time. GPT suggested I try highlighting next actions instead. It suggested I put a checklist in my Grant Application Trello card (duh) but this led me to the insight of putting the next action right there in the card title, which led to a whole different insight:
Specific action step first. Project second.
So instead of a Trello card reading: “Start OVC Services Grant,” the Title of the card now reads something like:
“Fill out grant revision matrix: OVC Services Grant”
This makes it so my brain registers the specific, small action step rather than the whole project at once.
I also worked on tweaking all my Trello cards in this format and operationalizing the task titles that felt vague or overly complex. So a task called “Look into HUD CoC grant” that I’ve been avoiding for literally months became “Find most recent NOFO: HUD CoC grant”
How hard can it be to find a NOFO?
Every time I complete an action step, I cross it off the card’s checklist and edit the title with the next action step. And so in this way I am only ever looking at the next step.
It’s bloody genius.
We worked on some other tweaks over the course of about four hours. What’s most impressive to me about this whole thing is that after I made these changes, I got three or four tasks done right away. I submitted a grant report, we created a revision matrix using the specific rejection feedback I got, we created a matrix template for future grants, and we created an Excel workbook to document general learnings from grant reviewer feedback.
It seems like the potential for integrating AI into the grants management workflow is pretty much unlimited, but right now I’m focused on developing what GPT calls my “working-memory-reduction-system.” Since I told it I have a tendency to hyperfocus on lower priority things, it told me to put working on my system in the Parking Lot until I complete my highest priority tasks.
I have deep existential thoughts, doubts and philosophical questions about this (and I think brain fry is a real thing after spending hours in constant dialogue with an LLM) but the bottom line is I think this project is going to help me a lot, in ways I didn’t expect.
