If you're juggling too many things, add this #SKILL to your AI model, thank us later..

It's been a busy few months for us folks over here at 3 Peat. As we grow and expand, our attention gets pulled in a multitude of directions. Each new AI related concept or business challenge presented for us to solve has the team back to the drawing board to determine whether said idea fits with our internal strategy.

I suspect your job and company is no different. There's always a shiny new toy being dangled in front of you, whether it be a new AI tool that will revolutionize your business strategy, or a consultant coming along proclaiming to automate half of your workforce (hah!).

Each idea, concept, opportunity, and challenge needs to be reviewed and considered thoughtfully. So how are we, as busy and important people, handling this these days? We throw it into an LLM!

"Claude, analyse this and give me the pros and cons" or "Copilot, what do you think of this concept?"

What you're doing is outsourcing triage. Perfectly normal and baked into our human nervous system. Given most people's time constraints, triaging is the ONLY way to get through a hectic day (as any parent of young kids knows; btw good luck for this coming school holidays).

Here's your top tip for today, and why reading this article will make it worth your while. If you're going to triage, please do it correctly.

Setup a #skill in your LLM. You can call it triage.md. What it does is follow a specific set of YOUR instructions when handling an idea or proposal.

So instead of just asking GPT "Hey what are your thoughts on this?" and letting your LLM go rogue depending on its mood, it will instead go "My owner is asking me for my thoughts on this. Before I start I should check the triage.md skill... it's telling me to consider the following:"

What triage.md should check

1. Has this idea been pitched before? Flag it if yes. They're very, very busy; they may have forgotten.
2. If it's new, review it against the knowledge base currently in the profile. Does it actually align with previous enquiries? Particularly if it is the opposite of something previously agreed upon.
3. Effort versus payoff. Rough cut only. High, medium, or low on both axes. Don't overthink it.
4. What decision does this actually require? Is it act now, park it for 90 days, or quietly never speak of it again?
5. Who needs to be involved? Does this need the founders, a client, or nobody at all right now?
6. Is there a vendor behind this pitch? Because if someone's trying to sell something, that context matters.

This is my triage.md. If you have a perfect memory, you can probably skip step #1.

The point of the skill though is to highlight a fundamental difference: Without triage.md, you're asking someone who's read every book ever written but has no idea what your actual priorities are. With it, you've handed them your own internal compass so it can take into account what's important to you.

And yes, I get the irony of it all. We're using AI to decide which ideas (most of which, also generated by AI) are worth pursuing. The machine is triaging proposals about machines!

But that's where we are. The volume of ideas, tools, consultants, and half-baked pitches isn't slowing down. If anything, it's accelerating. We've moved past the competitive advantage of access to AI. It's having a system that stops you from drowning in it.

triage.md: give it a shot and let me know if it's been helpful to you.