Rethinking PARA: why I added an operations layer

This post is the first in a series on OPERA, a framework I am developing that puts an automation layer on top of a human-centric filing system. It is based on my own setup, so treat it as one person's working notes rather than a recommendation.

From PARA to something with agents in it

I found PARA (Projects, Areas, Resources, Archives) several years ago and it reorganized how I keep my digital life, because the logic suited me: everything had a place, from laptop files to Gmail folders, Chrome bookmarks to Google Drive, and I ended up building a whole Notion setup around it. What changed is that AI tools arrived in the middle of that workflow, and I think a filing system designed for a person doing all the filing runs into limits once software can do some of it.

PARA, briefly

PARA is the most elegant knowledge and project management system I have used. Tiago Forte designed it around one premise: everything you work on fits into one of four categories.

1. Projects. Active, goal-driven efforts with a defined outcome, such as "Launch website redesign".
2. Areas. Ongoing responsibilities you manage over time, such as "Personal branding" or "Health and fitness".
3. Resources. Reference material that needs no action now but might be useful later, such as "Artificial intelligence" or "Productivity".
4. Archives. Inactive items that are no longer relevant but may need to be revisited.

What makes PARA work is that it adapts. A knowledge worker uses it to structure research, writing and projects; someone doing creative work uses it for inspiration, drafts and finished pieces; a student uses it for coursework and notes. It guides rather than dictates, which is why I am still using it.

The PARA structure in my Notion

Where it runs out of room

I rely on PARA and I still hit four problems, all of which got sharper once AI tools were part of the picture.

Information overload. I collect more notes, data references and articles than I have time to read, let alone analyze or use.

Curation. PARA keeps information organized, but organized is not the same as curated, and I suspect relevant material I already have is slipping past me unread.

Manual upkeep. Moving items between Projects, Areas and Archives takes time, and it becomes too much when several things are running at once.

No agentic support. With agent frameworks like CrewAI and AutoGen, platforms such as RelevanceAI and n8n, and coding assistants like Cursor, there is an obvious opening to extend PARA into automated task orchestration and knowledge curation, and nothing in the framework itself reaches for it.

None of this makes PARA a bad system; its simplicity is the reason it survived in my setup where more elaborate things did not. I want the automation built in, and as far as I know no formal framework offers that yet.

OPERA

OPERA (Operations, Projects, Environments, Resources, Archives) keeps PARA's structure and adds an Operations layer for task management, knowledge curation and automation.

Although I frame it against PARA, the problems are not specific to PARA. People running GTD or a setup they invented themselves report the same three things: too much incoming information, no structure that AI tools can act on, and task orchestration that stays manual. OPERA is a theoretical framework for now, aimed at those gaps.

What the layer is for

The Operations layer automates repetitive work, organizes knowledge, and surfaces what is relevant, working as an extension of what I do rather than a replacement for it. By "agentic" I mean AI tools that operate autonomously inside parameters I set, making decisions and adjusting workflows while I keep oversight. PARA's simplicity stays intact underneath.

The structure

Base layer (Projects, Environments, Resources, Archives). This is your normal workspace, whether that is Notion, Google Drive, or some combination. Each note or task carries a unique identifier so the Operations layer can reference it, which is the one real change from PARA: tagging good enough for software to act on.

• Projects. Goal-focused workstreams, as in PARA.
• Environments. The thematic contexts those projects live in, such as a coding environment or a design environment.
• Resources. A curated knowledge base that agents can both read from and add to.
• Archives. Long-term storage for completed or dormant items.

Operations layer (the agentic engine). An Orchestrator Agent assigns tasks to specialized crews, roughly the way an air traffic controller assigns runways, and it does more than delegate: it curates knowledge bases, retrieves and analyzes data, and keeps a workflow continuous across Notion, Chrome bookmarks, Gmail and Google Drive. The crews can be single agents with one clear job (a researcher), structured teams (a blog writing crew), or multi-step workflows spanning several teams (a stock analysis and trading strategy flow), built on frameworks such as CrewAI or AutoGPT.

The rest of the field

Plenty of companies are building in this direction. Relevance AI, Akira AI and Agentive all sell tools for exactly this combination of project management, knowledge curation and automation.

The large platforms are moving too. Notion, ChatGPT and Google keep adding AI-assisted writing, summarization and better search. If their native agent capabilities get strong enough, a framework like OPERA would need much less external automation, which would be a good outcome and not a threat to the idea.

Either way, I think the underlying principles of OPERA hold: knowledge you can retrieve, tasks with a structure, and AI that can act on both.

What I want to find out

I would like to know whether OPERA can actually carry simple delegated tasks, hold a piece of long research together, and put the right thing in front of me at the right time. I am going to build it in Notion and write up what happens, including the parts that do not work.