Where could “AI” be actually helpful?

It helped me revise some project outcomes today, and I learned stuff. The written outcomes I submitted were pretty shit and probably a core reason why the grant got rejected. Between GPT and a Grant Professionals Association federal grant-writing training I’m doing right now, I feel like I’m in a wonderful learning feedback loop where I can cross-reference and integrate my learnings.

Obviously I can Google the proper way to write objectives, but it’s useful to have a tool that can troubleshoot. In one case, we’re submitting a (previously rejected) federal grant under the priority of “combatting violent crime.” This is a victim services grant, so I explained to GPT that we don’t control the outcomes of prosecutions, etc. so I wasn’t sure how to approach promised outcomes. GPT gave me good feedback on how to adapt the original overly broad objective to a services-focused objective that connects survivors with resources to exercise their legal rights and pursue justice options.

I also asked it, like, hey, how do you handle it when your outcome is something you’re already doing? It gave me a pretty comprehensive response.

It’s like having a grant consultant to provide me unlimited on-the-job training. I am discovering that my work is actually more complex than even I realized and that there are countless ways I could improve my workflow. Keep in mind, I’ve been writing corporate and foundation grants for 16 years, with the occasional federal narrative, but I’ve only been managing federal grants for four years - and until recently, I never had any training. Because you have to write pages and pages of narrative, one could erroneously conclude that writing a federal grant is about, well, writing. But it’s much more about planning, development strategy, partnership outreach, project design, evaluation, etc. You could fill a book with the stuff I don’t know yet. And that’s where I think AI is going to have the most useful application: to teach me the stuff I don’t know, specific to my work, as I’m doing it.

They key thing for me is to have continued (actual expert) input through ongoing training to ensure what I’m learning from GPT is in line with standard practice. It’s the interplay between these two resources that I think is going to have the strongest impact.

Facebook has this new AI thing where there’s buttons to summarize any story that gets posted which basically kills clickbait headlines. I kinda like it to be honest.

Are you saying it is, or that you wish it is?

An AI bullshit filter would be incredibly useful. That’s a product I might pay money to subscribe to.

It is pretty useful, it completely shutdowns clickbait headlines by providing a small summary of whatever the headline is talking about.

Try it this way: “To be honest, I kinda like it.”

Today me & Claude did a deeper dive into my workflow issues, particularly with how my Trello board is set up and how I’m using it. I’ve found the changes it suggested before to be greatly helpful, but I still struggle with prioritization and getting lost in all those task cards.

It helped me better understand the purpose of each column and gave me a Prioritization cheat sheet for me to review every Monday. Basically it told me that once I move stuff into Focus This Week I have to stop looking at everything else until those tasks are done.

So I chose three tasks and that’s it. No rearranging my cards until those are done (unless something urgent comes up, but even then I can only remove one card from Focus this Week to make room for the new task.)

Boy is that anxiety provoking. I’m terrified I’m missing something. My whole life I’ve succeeded by obsessively checking and re-checking and reviewing what I have to do. The idea of just deciding and then trusting that it’s all there is really unsettling.

But all that checking and re-checking is a waste of time and mentally exhausting, so I’m giving this a shot. I get to review everything twice a week, Wednesdays and Mondays, and I have a cheat sheet for that process that feeds into my required Monday report to my supervisor.

Claude also gave me a learning program to create SharePoint automations to reduce my administrative burden when I submit or receive a grant. Right now I have to update half a dozen different locations which takes upwards of 30 minutes just to document a grant as awarded. I presented the problem to Claude, and apparently there’s a way to create a system where I input everything into one sheet and it auto-populates everywhere else. We talked about the pros and cons of SharePoint vs. Jotform and ultimately I chose SharePoint because I own the SharePoint space and wouldn’t be dependent on anyone to maintain it.

Claude helpfully told me the skills I would need to learn and where to find the tutorials. It’s quite a project but it feels doable. In the meantime, we made a checklist so none of the documentation steps fall through the cracks when I’m doing it all manually.

Claude also created a sheet for me to track outstanding funding needs in order of priority, reflecting what grants have been secured to fill each need. So I’ll be able to see at a glance what outstanding needs we have when considering what to apply for any given opportunity. For example, “we need to fund this staff position, this is priority 1, we have secured two grants of $3k and $20k toward this position and we have $25k remaining to be funded. Based on this information, I should write this grant request with $25k allocated to staff.”

I have yet to take this spreadsheet for a test run. But it looks promising (it’s also something I could never have organized so well on my own - that part of my brain is broken.)

Both of these are Parking Lot tasks - I’ll get to 'em when I get to 'em - but it’s very useful to have a start and a plan.