The AI System That Cut My 19-Hour Research Task to 6

Build your own judgment-driven system and use AI to deepen your insights, and stay true to your values.

Source: Diana

Last year, deep in writing about long-term memory, I realized I was losing mine.

Not literally, but functionally.

I’d scroll through AI tool directories, skim what felt like 200 “game-changing” launches a week, and test plugins that promised to “revolutionize research.”

Most left me mentally foggy and mildly annoyed. One gave me a perfectly structured summary of a paper it hadn’t actually read.

I was trying to write about cognitive integrity while outsourcing my cognition.

If you’ve ever opened five browser tabs trying to be more productive and ended up closing them all with nothing to show for it, you know the feeling.

So I asked a harder question: What would it take to use AI tools without trading away discernment, depth, or direction?

In this article, you will find my answer, not a list of tools (though I include the ones I continue to use), but a blueprint.

---

A Boundary System for Strategic Use (Your AI Research Charter)

Think of the Charter as your philosophical safety net and practical filter. It’s a living document to govern how and why you engage with AI tools.

Without one, every “assistant” risks becoming an attention thief in disguise.

Here are four dimensions mine includes:

1. Ethical Red Lines

What will you never let AI do on your behalf? Where does integrity take precedence over speed?

Mine: “I will never use AI to generate novel research claims without rigorous source verification and explicit attribution.”

Not because I’m afraid of being “caught,” but because the work loses meaning otherwise. Shortcutting trust collapses the very foundation I want to build on.

2. Mission Alignment

How does AI serve your goals, not just generic productivity?

Mine: “AI helps me synthesize multi-source research to sharpen insight frameworks, freeing time for original analysis and client impact.”

When AI pulls its weight, I write better, think deeper, and stay in the zone of meaning-making. If it’s just a fancier search bar, I’m out.

3. Cognitive Load Design

Where is your brain doing work that could be structurally offloaded — not outsourced entirely, but scaffolded?

Mine: “AI assists with broad literature reviews, first-pass theme coding, and structuring arguments but never leads with conclusion-drawing.”

This principle helps me stay present for the real thinking work. Not everything needs to be hard, but not everything should be easy either.

4. Quality Thresholds

How will you define “good enough” — not for AI, but for yourself?

Mine: “AI-generated drafts are a junior analyst’s first pass. Every key idea must be triangulated, challenged, and reworked before it becomes mine.”

These four dimensions changed the way I think and work. Here’s exactly how I built them into a workflow that saves 12+ hours per review—and the tools, prompts, and AI guardrails I rely on.

---

From Chaos to Cognitive Clarity (The Workflow)

Let’s walk through a real example: running a rapid evidence review on the efficacy of portfolio-based assessments for secondary school students.

The rest of this post is for paid subscribers.