- PM Network Alliance Newsletter
- Posts
- Cowork can now remember what you’ve already told Claude
Cowork can now remember what you’ve already told Claude
INCLUDED: A fun tool that helps you track, monitor, and manage your Claude usage limits

In Today’s edition:
Anthropic Gives Claude and Cowork a Shared Memory
PMNA TUTORIAL: How to Use Claude to Rescue a Messy Monday Board You Just Inherited
Salesforce Introduces Slack Code for Collaborative AI Coding for Teams
TOOL REVIEW: Diet Claude Helps You See Your Claude Usage Limits Before You Hit Them
Exciting Career Opportunities for Product and Project Management Professionals
Reading time: 5 minutes
HEADLINE NEWS
Anthropic Gives Claude and Cowork a Shared Memory
Anthropic has unified Claude’s memory system across its standard chat experience and Claude Cowork, allowing information learned in one workflow to carry over into the other. The update is designed to eliminate the need for users to repeatedly re-explain projects, preferences and other context when moving between Claude’s tools.
Shared memory: Claude can now use information from previous chats when starting a Cowork task, and context from Cowork can inform future conversations.
Real-time updates: Claude can generate and update memory topics during conversations instead of waiting until a chat ends.
More user control: Users can review, edit, delete, pause or reset saved memory topics through Claude’s settings.
Sensitive information is protected: Sensitive topics are excluded from memory by default, while users can separately opt in to storing some of them; certain categories remain excluded.
Broad availability: Memory is enabled by default for Free, Pro and Max users across web, desktop and mobile, while Enterprise and Teams users are excluded from the default rollout.
The change makes Claude feel more like a continuous assistant rather than separate tools that require users to repeatedly provide the same background information. At the same time, Anthropic is putting greater emphasis on memory controls, giving users ways to decide what Claude retains as that shared context becomes more deeply integrated into their workflows. Read More
ONE MONTH FREE OF MAKE.COM!
If you’ve been looking for a smarter way to automate your workflows, now’s the perfect time to try Make. It’s one of the most powerful tools for connecting apps, streamlining tasks, and building systems that run without you.
We have brought you a fine solution. This gives you full access to advanced features to enable you to automate lead gen, content distribution, or internal operations.
For a limited time, you can get 1 month of the Pro plan completely free using my link below:
PMNA TUTORIAL: How to Use Claude to Rescue a Messy Monday Board You Just Inherited
Inheriting a messy Monday board is a data-archaeology problem before it's a project-management problem; you can't fix what nobody has measured. Connecting Claude to Monday via MCP lets you audit the whole board numerically in minutes, then apply the corrections in verified batches. The workflow runs in five phases (audit, decide, remediate, lock, monitor) with every destructive step gated behind human approval and a full board snapshot.
Key Steps and Ideas:
Snapshot the board with
duplicate_board_with_pulses_and_updates— the other variants silently drop the comment history the audit reads.Confirm write permission before starting; some scopes fail even for board owners, and automation edits may need the monday UI.
Run a read-only audit: label distribution, ownerless items, overdue-and-open, dead columns, staleness, duplicates.
Audit the automations — the live ones are why cleanups quietly revert.
Make Claude propose two taxonomies with a full migration map; you pick one and lock it.
Split the map into renames and merges — rename labels in place with zero item writes, move items only for true merges.
Reassign the merge items in batches of 40 with
createLabelsIfMissing: falseso bad targets fail loudly.Verify the distribution matches the approved map — merged labels at zero, renamed labels holding their counts, total unchanged.
Flag duplicates for human review, then archive the losers — monday has no merge primitive, and archives are recoverable.
Write column descriptions — the cheapest, highest-leverage, most-skipped step.
Ship a triage view and a weekly health check saved as a monday doc, and read the deltas, not the numbers.
What You Should Consider:
Claude's value here is speed of diagnosis and precision of execution, not judgment — and its best moments are refusals: rejecting a verification premise that would have masked data loss, and declining to merge two duplicates when only a human knows whether the work shipped. Keep the human gate on the taxonomy and every archive decision, automate everything downstream of those, and the rescue holds. The board stays rescued only if the weekly check runs.
VIDEO WALKTHROUGH BELOW 👇
💡TUTORIAL VIDEO WALKTHROUGH 💡
Build a One-Session Cleanup for a Messy Monday CRM Board
This tutorial reflects and solves a common PM problem: Inheriting a messy Monday board. This is usually a data-archaeology problem before it's a project-management problem, and as always, you can't fix what nobody has measured. Connecting Claude to Monday via MCP lets you audit the whole board numerically in minutes, then apply the corrections in verified batches. The workflow runs in five phases (audit, decide, remediate, lock, monitor) with every destructive step gated behind human approval and a full board snapshot.
Adapted from today’s tutorial.
Salesforce Introduces Slack Code for Collaborative AI Coding for Teams
Salesforce has launched Slack Code, a new way for teams to work with AI coding agents directly inside Slack. Instead of developers working with AI in isolated tools or browser tabs, Slack Code creates dedicated project-based code channels where teams can collaborate with agents, review code and watch development happen in real time.
AI coding moves into Slack: Teams can work with Claude Code, Devin, GitHub Copilot, ChatGPT and Vercel agents from shared Slack code channels.
Everything stays in one workspace: Each code channel includes tabs for conversations, agent plans, code diffs and live previews, keeping the project context together.
Non-developers can participate: Product managers, designers and marketers can trigger agents, review their work and help guide development without needing to write the code themselves.
Human oversight remains central: Team members can pause, redirect or stop agents, while human approval is required before final changes reach production.
Available across Slack: Slack Code and code channels are available on all Slack plans, although individual AI-agent integrations may require separate access to those services.
Slack Code represents Salesforce's push to turn AI-assisted software development from an individual workflow into a team activity. By bringing agents, conversations, code reviews, and live previews into a shared environment, Salesforce aims to reduce context switching while making AI-generated development work more visible, collaborative, and governed. Read More
TOOL REVIEW: Diet Claude Helps to Track Your Claude Usage Limits Before You Hit Them
Diet Claude is a free Chrome extension that surfaces your Claude usage as a live meter; a soda can that drains through your 5-hour session and weekly window, then refills at reset. It reads the usage figures Claude already reports for your own account, in your own browser, with no account or API key to set up. Alongside the meter, it bundles tools for stretching whatever quota you have left and for carrying work elsewhere when you run dry.
Live usage meter — shows how much of your 5-hour session and weekly window you've spent, as a draining can or a numeric panel.
Reset countdown and threshold alerts — desktop warnings at 70% and 90%, a ping when the session resets, and a one-hour heads-up while usage remains.
Context gauge and cache timer — a "tokens used" readout on the message box plus a warmth window showing when a reply still reuses the chat cheaply.
Fork & Compress — summarizes a heavy conversation and continues it in a fresh chat, keeping the gist and dropping the expensive history.
Continue in another AI — one click summarizes the open chat and types it into ChatGPT, Gemini, or Grok, ready for you to review before sending.
If you run Claude-driven monitoring, rollups, or reconciliation loops, your quota is a scheduling constraint; and right now an invisible one that strands workflows mid-run and forces restarts you didn't budget for. Diet Claude makes it legible before it bites: you can see whether a long run fits the window you have left, sequence heavy work against the reset clock, and cut context cost on expensive chats.
For anyone treating Claude as operational infrastructure, that's the difference between planning around a limit and being ambushed by it.
Exciting Career Opportunities for Product and Project Management Professionals
R&D Project Manager CS US @ Cybex
📍 Charlotte, NC.
Sr. Product Manager - Compliance Document @ Avalara
📍 Durham, NC.
Senior Project Manager, Strategy and Execution @ FUJIFILM
📍 Holly Springs, NC.
Product Mgr II @ MetLife
📍North Carolina.
Project Manager - Nuclear @ Sargent & Lundy
📍 Southport, NC.
THAT’S A WRAP
Thank you for being a part of our growing community. We look forward to sharing valuable content, industry trends, and strategies that will help you navigate and lead in this dynamic space. Stay tuned for more to come! Best, | ![]() |
