Claude assistant
Claude, wired into your workspace
It reads what you can read and does what you can do. With your permissions, in your workspace, and nowhere else.
The panel is your agent with your workspace in front of it. Claude, Codex or Cursor’s, whichever you connect. It can read records, conversations, meetings and reports, and it can act: create and update records, send and schedule messages, build sequences, run reports, change settings. With your permissions, in your workspace, and nowhere else.
What it is good at
- Questions across your data
- What happened with them this quarter. Who has not been touched in thirty days. Which template gets replies.
- Multi-step work
- Import these twenty people, find their emails, enroll them in the outbound sequence.
- Building things
- Sequences, workflows, templates, ICPs, reports, forms, docs.
- Explaining the product
- Ask how do I about any feature, and it answers, or it does the setup with you.
What it can see
The page you are on, which is what makes this person resolve to the person on screen. Anything you paste or attach. And through its tools, everything in the workspace your permissions allow.
Where it lives
- The panel in the app, on every page. ⌥Space opens it, Alt+Shift+Space on Windows.
- The desktop app, where it runs locally and picks up the same conversation.
- Claude Desktop, Claude Code, Cursor or any MCP client, through the MCP integration.
- On a clock, as a routine. On a card, as a skill.
Memory
What you tell it to remember, it remembers everywhere: the panel, the desktop app, claude.ai. How you like a follow-up written. Who a person is to you. What a project is for.
Getting more out of it
- Ask for the outcome, not the steps. Get me a meeting with someone there beats a list of micro-instructions.
- On a bulk edit, ask it to check its own work afterwards.
- Anything you ask repeatedly, save as a skill.
- Give it a page. On a person, this person means them.
Around claude assistant
The parts of the workspace this one works with.
Skills
Saved prompts as runnable cards. Describe the recurring ask once, then it is one click.
Anything you find yourself retyping into the assistant is a skill waiting to be saved. A skill is a card with a name, a prompt, the tools it may use, and a Run button.
Routines
The assistant on a schedule. Every Monday at eight, summarise last week’s pipeline and post it.
A routine is a prompt with a clock. It runs in its own chat, with your permissions, and tells you when it is done. Anything you would ask the assistant every week is a routine waiting to be written.
Claude Desktop and MCP
Your workspace as tools inside Claude Desktop, Claude Code, or any MCP client. Same permissions, same memory.
Everything is an MCP server. Connect it and every record, conversation, meeting and report is a tool the model can call from wherever you already talk to it: Claude Desktop, Claude Code, Cursor, ChatGPT, or an agent you wrote.
The setup checkup
Ask the assistant what is connected and what is not. It audits the workspace and walks you through the gaps in order.
Ask what should I set up, and the assistant checks the workspace with its own tools and comes back with the two or three highest-impact gaps, not a checklist of twelve. Then it fixes what it can.
Questions
Can it send mail without asking?+
It can, with your permissions. Ask it to draft and you review, or let it send and it tells you what went out.
Does it see other people’s data?+
Only what your permission profile allows. A member’s assistant sees a member’s view.
Which agent?+
Whichever you connect: Claude, Codex, or Cursor’s. It runs in the app, or locally through the desktop app when one is awake.
How is a skill different from a routine?+
A skill runs when you click it. A routine runs on a clock. A routine can run a skill.
Can I use a skill from Claude Desktop?+
Yes. Skills are listed to any connected MCP client.
Can a routine send email?+
Yes, with your permissions. Most people have it draft and let the brief show the drafts for approval.
What if it fails?+
The run is marked failed with the reason, and you are notified.
Is this the API?+
It is the API for agents. For a script, the same tools can be called over HTTP with a token from the panel.
Can I limit which tools a client gets?+
Yes. A connection can be scoped to read only, or to a set of tools.
Does the checkup run on its own?+
No. Ask for it whenever you want a read: after inviting the team, after a migration, or when something feels missing.
Can a teammate run it?+
Yes. It reads with their permissions, so an admin sees more gaps than a member.
Try claude assistant on your own data
Connect your email, then open a person. Most of the product shows up on that one page once the mail is in.
More in Automate
Sequences, workflows, forms and an assistant that does the repetitive parts.
A reply arrives, a deal changes stage, a form is submitted. A condition decides, and the workspace does what you would have done by hand.
A block-based builder. A submission creates or updates a person and can trigger a workflow.
Describe the recurring ask once, then it is one click.
Every Monday at eight, summarise last week’s pipeline and post it.