Your AI is forgetting your context (mine was too)

Hey, Yunus here 👋

The last few weeks weren’t quiet because nothing happened. They were quiet because I finally tackled a problem that has been slowing me down for months. A problem you may have run into too.

AI kept losing my context.

Not because the tools are bad, but because I rely on them across too many surfaces: the apps I’m building, digital products, client work, experiments, and this newsletter. Each assistant only understood a fragment. None of them could hold the full picture. And since most of you here are indie hackers or solo creators, I know you wear multiple hats every day. You’re switching contexts constantly. So this isn’t just my problem; it’s very likely yours too.

So I started building a structure that gives AI a single, unified brain to operate from.
Here’s what I actually did:

PS: If you want me to share the full setup, reply with SYSTEM. I’ll consider turning it into a simple, easy-to-use framework.

1. I migrated everything into a PARA-based folder system

→ Projects, Areas, Resources, and Archives.
→ Simple, but grounding.
→ Every ongoing project has a real home now.

2. I created a file called business_map.json

This became the “core brain” layer of my entire operation.
It’s not a task list. It’s a structured map of how my business actually works.

Instead of tracking to-dos, it captures the deeper layers that AI usually misses:

• the legal and financial structure

• how money flows across mobile apps, info products, services, and hosting

• monthly revenue patterns and insights

• the full project (and products) portfolio (mobile apps, web products, info products, services, media brand, customers, etc…)

• each project's status, tech stack, constraints, priorities, and recent achievements

• a brain hub with ideas, active focus zones, sprints, and knowledge areas

In other words, it’s the unified story AI never had:
what I run, how it fits together, what matters now, and what is on hold.

Now, when AI helps me, it can navigate meaning instead of isolated tasks.
It can understand the difference between a mobile app sprint, a freelance client, a digital product, or a newsletter goal, because the map explains the context behind each.

Here is a blank starter template if you want to explore the structure without any private details:
→ View business_map.json (starter template)

3. Each project got its own project_context.md

Inside each project folder I wrote down only what actually matters:
why the project exists, what I’m trying to achieve, the constraints, the blockers, and the key notes that shape decisions.

For example, in Inbox Experiments, I keep both brand_guide.md and old_issues.md inside the project folder.
This way, whenever I ask AI for help about the newsletter, it has the full context, tone, strategy, past issues, and current direction. At minimum, it knows as much as I do.

No more re-explaining the same story to AI over and over.

The outcome

Once AI had a unified structure to work from, things shifted immediately:

• Better continuity

• Fewer resets

• More relevant answers

• Decisions felt connected instead of isolated

And most importantly, I didn’t need to “remind” the assistant who I was or what I was doing every day.

I’m curious about you.

Do you use anything like this?
A system, a file, a note, a folder structure that helps AI keep track of your world?

If you want, I can go deeper and share the entire structure I’m using.
Just hit reply and tell me you want it.

Talk soon,
Yunus