Workplace Memory

Last verified: 14 April 2026 | Applies to: Pro, Max, Team, Enterprise

In 30 seconds

Claude has a built-in Memory feature (Settings → Capabilities) that passively learns your name, role, and preferences from conversations over time. It’s useful, but it’s automatic, you can’t control exactly what it stores, and it only captures what comes up naturally in chat. The Productivity plugin gives you a deeper, structured layer on top: you seed your full business context explicitly (team names, acronyms, processes, priorities) and Claude loads it every session. Set it up once (about fifteen minutes), and every Cowork session starts with Claude already briefed. This is the first workflow most operators should complete.

Prerequisites

• Claude Desktop installed (download)
• A paid plan (Pro or above)
• The Productivity plugin installed (see Step 1 below)

Step by step

1. Install the Productivity plugin

   Open Claude Desktop → Cowork tab → Plugins (in the left sidebar) → search “Productivity” → Install.

   You can also install from claude.com/plugins.

2. Initialise your workspace

   Open a new Cowork session and type:

   /productivity:start

   Claude creates four things in your working folder:

   • TASKS.md: a task list Claude can read and update
   • CLAUDE.md: working memory for your current context
   • memory/: a directory for longer-term knowledge
   • dashboard.html: a visual overview that opens in your browser

3. Seed your context

   Claude will ask about your role, team, and current priorities. Be specific. The more you give, the less you’ll have to repeat later. Here’s an example:

   I'm COO at a 15-person ecommerce company. My direct reports are Sarah (ops), Marcus (finance), and Priya (marketing). We're mid-way through migrating from Shopify to a headless setup. Key acronyms: PSR = product status report, WBR = weekly business review, Q1 close = our fiscal year ends in March.

   This gets stored in the memory files. Next session, Claude already knows.

   What to include:

   • Your role and company (size, industry, stage)
   • Direct reports and key stakeholders (names, roles)
   • Current projects and priorities
   • Acronyms and shorthand your team uses
   • Preferences (report format, communication style, tools you use)

4. Test the memory with shorthand

   Once your context is seeded, test it:

   Draft a message to Marcus asking him to prep the PSR for tomorrow's WBR

   Claude should know that Marcus is your finance lead, what a PSR is, and that WBR means your weekly business review. No clarifying questions needed.

   Try a few more:

   • “What are my current priorities?”
   • “Create a task: review Q1 close timeline with Marcus by Friday”
   • “Open my dashboard”

5. Keep it current

   Run /productivity:update weekly to triage stale tasks and flag memory gaps. This takes about two minutes.

   Use /productivity:update --comprehensive monthly for a deeper scan. Claude reviews all stored context and suggests what’s outdated.

   The rhythm: Seed once, use /productivity:update weekly, /productivity:update --comprehensive monthly. Total maintenance: five minutes a week.

How operators actually use it

Scenario: James, Founder/CEO

James runs a 12-person B2B SaaS company and spent about fifteen minutes seeding his workplace memory — team structure, current projects, key acronyms, and his preferred reporting format. The following week, every Cowork session started with Claude already knowing his direct reports by name, understanding that “PSR” means product status report, and recognising the company’s fiscal calendar. No more re-explaining context at the top of each conversation. James estimates the setup saves him five to ten minutes per session, which across three or four daily sessions adds up to a couple of hours a week.

Built-in Memory reduces the cold-start problem in Chat, but it’s patchy. It may know your name and role but miss your team’s acronyms or reporting preferences. Workplace Memory eliminates the gap entirely for Cowork sessions. You open Cowork and start working. Claude knows your team, your projects, your terminology. Twenty minutes of initial setup saves hours per week.

Built-in Memory vs. Workplace Memory

graph TD
    Chat[Chat conversations] --> Builtin[Built-in Memory]
    Builtin --> Passive[Passive context in Chat]
    You[You seed context] --> Productivity[Productivity plugin]
    Productivity --> Files[CLAUDE.md + memory files]
    Files --> Cowork[Structured context in Cowork]
    Passive --> Combined[Claude knows your world]
    Cowork --> Combined

Claude now has two layers of persistent context. They’re complementary, and most operators should use both.

	Built-in Memory	Workplace Memory (this workflow)
How it works	Claude automatically synthesises summaries from your chat history every ~24 hours	You explicitly seed context via structured markdown files
Control	Passive. Claude decides what to remember.	Full control. You write exactly what Claude should know.
Where it works	Chat, Desktop app, Mobile	Cowork sessions only
Scope	Per-project (each Project gets isolated memory)	Per-folder (lives in the directory you choose)
What it stores	Your role, preferences, working style, project details	Team structure, acronyms, processes, priorities, task lists
Editing	View and edit the summary in Settings, import/export supported	Edit the markdown files directly, any time
Plans	All plans including Free	Pro, Max, Team, Enterprise (requires Productivity plugin)

The practical difference: Built-in Memory is a colleague who gradually learns your habits from working with you. Workplace Memory is handing that colleague a detailed briefing document on day one.

Our recommendation: Turn on built-in Memory (Settings → Capabilities) for passive context in everyday Chat. Then set up Workplace Memory below for structured work sessions in Cowork where you need Claude to know your exact team, terminology, and processes.

Folder instructions (the lightweight alternative)

If you don’t want the full Productivity plugin, you can get basic persistent context by creating a CLAUDE.md file in any folder where you work with Cowork.

# Company Context

## Team
- CEO: Alex
- COO: You
- CFO: Marcus
- Head of Marketing: Priya

## Current Priorities
1. Website redesign (launch: 15 March)
2. Q1 financial close
3. New hire onboarding (2 positions)

## Terminology
- PSR = Product Status Report
- WBR = Weekly Business Review
- P0 = Critical priority

Cowork reads this file automatically when working in that directory. It’s less powerful than the Productivity plugin (no task management, no dashboard, no /productivity:update command), but it’s a good starting point if you want something simpler.

You can set global instructions (apply to all Cowork sessions) or folder-specific instructions (apply only in that directory). Folder instructions override global ones when they conflict.

What to watch out for

• Memory is folder-scoped. The Productivity plugin’s memory lives in the folder you selected. If you start Cowork in a different folder, Claude won’t see the memory from the first one.
• Cowork usage is higher than chat. Setting up memory uses Cowork, which consumes more of your plan’s usage than standard chat. The initial setup is worth it, but be aware.
• Keep context concise. Claude works better with structured, bullet-point context than with long paragraphs. Be specific, not verbose.
• Review stored context periodically. Team members leave, projects end, priorities shift. The /productivity:update --comprehensive command helps, but a manual review every quarter is worthwhile.

Related

• Plugins Overview: understand what plugins are and how they work
• Your First 30 Minutes: this workflow is part of the setup guide
• Cowork: understanding the environment where plugins run
• Skills: how Claude’s knowledge layer works

---

Something wrong or outdated? Let us know →

Get weekly workflows. Subscribe to the newsletter.