All jobs

Draft a Marketing Agency's Weekly Client Update From a monday.com Board

A marketing agency's agent reads one client's monday.com board each Friday and drafts the weekly update: done, in progress, waiting on the client, and due next.

Run this on RushOpens the console with this prompt ready to send. Connect monday.com there first.

What the agent does

This is for a marketing or creative agency that runs each client's work on its own monday.com board: one item per deliverable (a landing page, a set of ads, a newsletter), with a status, an owner, a due date, and the comments the team leaves as updates. Every Friday each account manager writes the client a summary of the week, mostly by scrolling the board. The run names monday.com's remote MCP server, connected once with POST /me/connections {"type": "mcp", "name": "monday", "url": "https://mcp.monday.com/mcp", "auth": {"mode": "oauth"}} and approved on monday.com's own page.

The agent reads the board's shape first with get_board_info, which returns its columns, groups and owners, so it knows which column is the status, which is the due date and what the agency calls each status label. Then it pages through every item with get_board_items_page, and reads the week's comments with get_updates and the week's changes with get_board_activity, which show what moved from "Working on it" to "Done" and what was added since last Friday.

It sorts the deliverables into four parts. Done this week: items whose status became done in the last seven days, by the activity log, not by guessing from the current status. In progress: items being worked on, with their due date. Waiting on you: items in the status the agency uses for client feedback or assets, each with what is needed, taken from the latest comment. Due next week: items with a due date in the next seven days. It also writes a separate internal note for the account manager, never for the client: items past their due date, items with no owner, and items that have not changed in two weeks.

The client-facing draft is written in the agency's voice from the template in its input, names deliverables as the board names them, and states no date the board does not show. The agent creates and changes nothing on the board.

Start it with one API call

POST https://api.phoenix.so/o/{org}/sessions with this body:

{
  "agent": "claude",
  "input": "Use the monday connection. Read the board named 'Northwind - Q4 campaigns': first get_board_info for its columns and status labels, then every item with get_board_items_page, the last 7 days of updates with get_updates, and the last 7 days of activity with get_board_activity. Write update.md for the client with four sections: Done this week (status changed to Done in the last 7 days, per the activity log), In progress (with due dates), Waiting on you (items in status 'Client review' or 'Needs assets', each with what is needed from the latest update), and Due next week. Use this opening: 'Hi team, here is where things stand this week.' Write internal.md for the account manager only: overdue items, items with no owner, and items unchanged for 14 days. State no date the board does not show. Create and change nothing on the board.",
  "connections": [
    "monday"
  ]
}
  1. Get an API key. In the console, open Settings, then API keys, and create a key for your org. Send it in the Authorization header as a Bearer token.
  2. Send the request. POST the body in request.json to https://api.phoenix.so/o/{org}/sessions, with your repository as the resources URL. The run starts in its own cloud sandbox.
  3. Read what came back. Follow the run on its live stream or wait for the run.completed webhook, then read the result and its trajectory.

What comes back

The session returns two files. update.md is the draft for the client, short enough to paste into an email or a client portal, with the four sections in a fixed order so clients learn where to look. internal.md stays inside the agency: the overdue, unowned and stale items the account manager should sort out before they hit send.

The account manager edits the draft before it goes out. They check that "done" items really are done from the client's side (a page can be marked done internally before it is live), soften or remove anything sensitive in the comments the agent summarized, and add whatever happened off the board, such as a call. The transcript keeps every monday.com call and its answer, so a line in the draft can be traced to the item and comment it came from.

Run it on every Friday afternoon

Run one per client board each Friday. In the console, a Project automation with a schedule trigger (0 13 * * 5 in the agency's timezone) can run a saved agent that names the monday connection, with the client's board name in its prompt; an agency with many clients gives each board its own automation, or starts the runs from its own tools with POST /o/{org}/sessions, one per board.

With a webhook URL on the API key, each run.completed tells the agency's handler which board's draft is ready, so it can notify that account manager directly.

The use case this job follows, with its automation and trigger: Produce a report or file, no repo.

Why run it on Rush?

The weekly update is the same reading job for every client, and it is easy to miss what changed when you only see the board's current state. The agent reads the activity log for what moved, writes the client part and the internal part separately, and leaves a record of every item it looked at.

The monday.com sign-in stays in a Rush connection rather than in a script's token, and each client's run gets its own computer, so every board's draft on a Friday runs side by side instead of filling each account manager's afternoon. Each run's cost is on its session, which an agency that bills by client can add up per board.

Frequently asked questions

Does it post the update to the client?

No. It writes a draft file and changes nothing on the board. The account manager edits and sends it, because the client relationship is theirs.

Does it use up our monday.com API limits?

monday.com runs each MCP tool call as a GraphQL API request, and those count toward the account's daily API call limit. One board's weekly read is a handful of calls, but it is worth knowing if the agency runs many boards often.

Our status labels are different. Does that matter?

Name your labels in the prompt. The agent reads the board's columns and labels with get_board_info before anything else, so it works from what the board actually uses.

Can it post the draft as an update on the board for review?

monday.com's server has create_update, which posts a comment on an item, so a prompt could ask for that on an item the agency keeps for drafts. This prompt keeps the run read only.

More monday.com jobs

More jobs like this: Produce a report or file, no repo

Keep reading