Yes. An AI agent can automate monday.com status updates: it reads items, status history and update threads, then posts a rollup on a schedule to a pinned item, a Slack channel or a drafted email. Statuses, due dates and anything a client reads stay with a person.

This guide covers seven ways to do it as of October 2026, with costs, limits, API calls and a worked Friday routine. I read monday.com's pricing page, its developer docs and its support articles on 9 October 2026. I have not run these setups on a live account, so I report what each vendor says.

In my AI agent demand by profession index (data pulled 4 October 2026), project managers rank 11th of 137 professions by US search demand for AI agents and AI tools, at 1,021 searches a month.

Key takeaways

  • Native automations post on a timer but send text you wrote in advance; they do not summarise the board.
  • AI Blocks cost eight credits per trigger, one item at a time. Weekly per-item summaries on 180 items use about 6,235 credits a month, about twice a Pro account's 3,000.
  • Through monday MCP, Claude, ChatGPT, Copilot or Cursor can reach your boards; whether a rollup recurs depends on the assistant you connect. Stiltsoft says sidekick does not handle scheduled delivery.
  • In my worked example, an external agent makes about 10 to 20 API calls for a weekly three-board rollup; Standard allows 1,000 a day.
  • Give the agent read access and update-write scope only. People change statuses and dates, and a person approves anything a client sees.
Four panels of board status bars feeding thin lines into one glowing saffron tray
Many boards, one weekly rollup.

Seven ways to automate status updates in monday.com (October 2026)

Option What it does Runs on a schedule? Cost or limit (checked 9 Oct 2026)
Native automations Sends a message you wrote in advance Yes Automation actions a month: Standard 250, Pro 25,000
AI Blocks (Summarize text) Summarises one item's text or update thread Runs per item, each time its column or automation fires Eight credits per trigger; AI columns on Pro and Enterprise
monday sidekick Summarises updates and conversations No (per Stiltsoft) Lite: up to 5 messages or actions per member a day
monday's agents (Risk Analyzer) Detects schedule, dependency and workload risks Watches "in real time"; no written rollup described Credits per agent run; no fixed number published
monday MCP with Claude or ChatGPT Reads and writes what you ask Depends on the assistant you connect No additional cost; calls count toward the daily API limit
Board Email Reports (Stiltsoft) Emails board reports Yes, via Custom Reports Price not in my sources; outside the daily API limit
External agent on the API Reads items, history and threads; writes the rollup Yes 1,000 calls a day on Standard, 10,000 on Pro, plus the agent's price

Native automations: fixed reminders on a timer

Recurring recipes start with an "Every time period" trigger, such as a reminder to a team every Monday at 9am. Time-based automations run in the desktop time zone of whoever set them up. The real limit: they send what you wrote in advance; they do not summarise the board. Plans cap them: Standard includes 250 automation actions and 250 integration actions a month, Pro 25,000 of each, and Enterprise 250,000 automation and integration actions combined.

AI Blocks and the Summarize text action

Summarize text is one of several AI blocks, usable in columns, automations and the workflow builder. Per monday support, "Triggering an action will use eight credits." Each trigger works on one item's text or update thread. Developers get configure_summarize_ai_column for an item's update thread, but only from API version 2026-10, still the release candidate while 2026-07 is current.

The built-in assistant: monday sidekick

The product page says sidekick can "summarize updates and conversations". The allowance depends on the tier:

Stiltsoft added a "Send Email" skill so sidekick can email a report on request. Stiltsoft also says sidekick is meant for on-demand reports and does not handle scheduled delivery (Stiltsoft makes the scheduled alternative, so it is not a neutral source).

Risk Analyzer and the other monday agents

Risk Analyzer "Detects schedule, dependency, and workload risks across projects in real time". No agent on monday's agents page is described as a status-reporting agent (as of 9 October 2026). Agents draw credits per run, with no fixed per-run number published; my explainer on monday's own agents and their credit pricing goes deeper.

Claude or ChatGPT through monday MCP

Through monday MCP, Claude, Microsoft Copilot, ChatGPT and Cursor can work on your boards, on all plans at no additional cost, within the connected user's permissions. Whether a rollup recurs depends on the assistant you connect, not on monday.

A marketplace app: Board Email Reports

Stiltsoft's Board Email Reports has a Custom Reports feature for recurring reports, the job sidekick leaves undone. My sources carry no price for it, so check the listing.

An external agent on the GraphQL API

An agent outside monday.com should connect with the narrowest access that still lets it read boards and post updates: an app scoped to updates:read and updates:write plus board reads, or the most limited user your plan allows. It reads items, activity logs and update threads, and posts with create_update, a Slack message or a drafted email. Swiftask advertises one such integration; its example runs every Friday at 4 PM, scans boards and sends a summary by email, Teams or Slack. Because it reads history and threads, it can say why an item is late.

The short version. Native automations run on a timer but do not summarise the board. AI Blocks work one item at a time, sidekick works on request, and MCP depends on the assistant you connect. A whole-board rollup on a schedule needs a marketplace app or an external agent.

The credit arithmetic: why per-item AI summaries get expensive

Here is my arithmetic on monday's published figures, for three boards with 180 items:

  1. Summarize text uses eight credits per trigger.
  2. Run it on every item once a week: 180 x 8 = 1,440 credits per run.
  3. About 4.33 weeks a month: 1,440 x 4.33 = about 6,235 credits a month.
  4. At $0.01 a credit (yearly billing), about $62 a month. At $0.0125 (monthly billing), about $78.

A Pro plan on monday.com includes 3,000 AI credits a month, and the pool is account-wide: "Your entire organization shares one credit bucket regardless of how many people are on the account." Per-item summaries alone would use that pool about twice over. Beyond those 3,000 included credits, the extra 3,235 come to about $32 a month at the yearly-billing rate, about $40 at the monthly rate. Agent runs draw from the same bucket.

One rollup per board per week is a handful of API calls. 180 per-item summaries a week is a budget line.

What a good status update contains

A good rollup answers six questions, in the same order every week so readers learn where to look.

Team version, grouped by owner

Each person sees their own overdue items, stuck items and deadlines. My guide to an agent that collects Slack standups covers the input side of the same loop.

Five lines for a stakeholder or client

The same headings, with overdue and stuck merged into one risks line and no owner names (a client does not need to know who is late). It always waits for a human.

Five board-health checks worth automating

Each check is a pattern a manager misses once they run several teams. Thresholds belong to your team, set per board; the numbers below are defaults you can change.

  1. No owner. The item sits in an active group with an empty person column.
  2. No status movement in N working days. Your team picks N; an engineering board and a content board move at different speeds.
  3. Suspected blocked. A blocked-style label, a dependency not yet Done, or an update that says "blocked", "waiting on" or "depends on".
  4. Due date pushed three or more times. Three is a default. The agent shows the pattern. The lead decides whether the item is uncertain or quietly dropped.
  5. Unanswered question older than five working days. Someone asked in the update thread and nobody replied. Five days is a default too.

How an agent reads and writes monday.com

The API is GraphQL. Three queries and one mutation do the work.

Reading items with items_page

items_page returns up to 500 items per page (default 25) with cursor pagination. Its query_params filter by column rules: any_of, not_any_of, between, is_not_empty, contains_text.

History from activity_logs

The item shows today's value. The log shows how it got there. activity_logs sits inside a boards query and filters by from and to timestamps, column_ids, item_ids, group_ids and user_ids, up to 10,000 logs per query.

Reading and posting updates

Reading updates returns body (HTML), text_body (plain text), created_at and creator, up to 100 per page. create_update(item_id, body, parent_id, mentions_list) posts an update or a reply and can mention a user, team or board. The scopes are updates:read and updates:write.

Versions and daily limits

Pin a version. As of 9 October 2026, 2026-07 has been current since 1 July 2026, 2026-10 is the release candidate and 2026-04 is in maintenance; new versions arrive quarterly. The daily limit resets at midnight UTC and covers personal tokens and private and public apps, plus calls through monday's hosted MCP server. Go over and you get DAILY_LIMIT_EXCEEDED.

Plan Seat price API calls a day Queries a minute AI credits a month
Free$0 (up to 2 seats)1,0001,000None (AI starts at Basic)
Basic$91,0001,0001,000
Standard$121,0001,0002,000
Pro$1910,0002,5003,000
EnterpriseCustom25,0005,000Custom

Seat prices are per seat a month, billed annually; paid plans start from 3 seats.

Worked example: the Friday 15:00 rollup

This is the routine I would set up. It is illustrative: the boards and numbers are invented, and none of it comes from a customer. Three boards (Website relaunch, Q4 campaigns, Client onboarding), 180 items in total, a Standard or Pro plan, and a summary board with a pinned "Weekly status" item. The agent runs every Friday at 15:00. The project lead reads it at 15:30, before anything goes to the client.

  1. Read the items. For each board, items_page at up to 500 a page, so one page each: name, group, status, owner, due date, timeline.
  2. Read the history. This week's activity_logs for the status and date columns (Monday 00:00 to Friday 15:00) show what changed. For flagged items, the agent reads that item's status and date history in activity_logs (filtered by item_ids and column_ids) to see how long it has been Stuck and how often its due date moved.
  3. Read the week's updates. Scan text_body for "blocked", "waiting on" and questions with no reply.
  4. Compute. Done this week, overdue, stuck more than three working days and due next week, plus the five board-health checks above.
  5. Write two versions. The team version by owner, the stakeholder version in five lines.
  6. Post. create_update on the pinned item and a message to #project-updates (see my guide on connecting an agent to Slack). The client email waits as a draft for the lead's approval. One comment per stuck item, mentioning the owner, at most once a week.
  7. Stay inside the budget. About 10 to 20 API calls per run: 3 item reads, 3 activity-log reads, up to 3 history reads, a few update pages, and 6 writes this week (the rollup plus one comment per stuck item). Standard allows 1,000 a day; Pro allows 10,000.

Say the week ends like this (invented numbers): 180 items, 112 Done (62%), 9 overdue with 4 of them on one owner, 5 stuck more than three working days, 14 due next week, 2 decisions needed from the client. The stakeholder version leaves owners out:

Weekly status, Friday (illustrative)

  • Headline: at risk; delivery holds if the two decisions below land by Tuesday.
  • Progress: 112 of 180 items done (62%) across all three workstreams.
  • Risks: 9 items overdue and 5 stuck for more than three working days, each with a named next step.
  • Next week: 14 items due.
  • From you: approve the homepage copy by Tuesday, and confirm the Q4 email launch date.

At 15:30 the lead reads it, fixes a line and approves the email. The run uses at most 2% of Standard's daily calls and no monday AI credits, because the agent writes the summary itself.

Guardrails for a status agent

A status agent needs write access to post its rollup, so I would hold it to five rules.

Why so strict? My AI agent incidents database logged 63 public incidents as of 6 October 2026. Twelve were a wrong commitment or action and nine a destructive action. That is a third of the total, and an agent that moved a due date on its own would add to the first of those two categories.

Connect the right user. Assistants on monday MCP can only do what the connected user is already permitted to do, so an admin connection gives an admin's reach. Risk Analyzer's page lists "reassigning owners, updating timelines" as mitigation; for reporting, I would keep that power with a person.

Should you automate this at all?

Automate it when the rollup spans more than one board or goes to a client every week. Keep it by hand when one person reads a rollup for one board; an agent adds setup and review for little time saved.

My free can-AI-do-this-task checker takes a task you do by hand, for example "weekly status rollup", and shows how much an agent could take on, how often the task comes around and what your time on it costs a month, with no sign-up.

Where to start

Start with one board and one audience. Pick the board people ask you about most, write the rollup you would want, and run the agent beside it in dry run. Add the stakeholder version and the nudges once the two match. My guide to AI agents for project managers covers the rest of the job.

Where Gravity fits

I build Gravity. It is in public beta. Its agents are already built: you describe the task in plain words, with no workflow to build. The catalogue is still short, so do not assume it covers your board routine yet. Gravity's Free plan is $0, Gravity's Mini Pilot plan is $15 a month (one automated agent, usage included) and Auto Pilot is $20 a month, with no seat pricing. If that fits, Explore agents.

Frequently asked questions

Can monday.com write status updates automatically on its own?

Partly. A recurring automation posts on a schedule, but it sends text you wrote in advance. AI Blocks summarise one item at a time, sidekick works on request (per Stiltsoft), and with MCP, whether anything recurs depends on the assistant you connect. A rollup that reads the whole board on a schedule needs a marketplace app or an external agent.

Can an AI agent post updates to monday.com?

Yes. The create_update mutation posts an update or a reply on an item and can mention a user, team or board. The agent needs the updates:write scope and nothing more on the write side.

How does an agent flag overdue and stuck items?

It reads every item with items_page: past due and not Done means overdue. For stuck items it reads that item's status history in activity_logs to see when it went to Stuck. Anything past your team's threshold goes into the rollup, and the owner of each stuck item gets one comment that week.

Does monday sidekick do scheduled reports?

Not according to Stiltsoft, which says sidekick is meant for on-demand reports. Stiltsoft points to its own Custom Reports for recurring ones.

How many API calls does a weekly rollup use?

In my worked example of three boards and 180 items, about 10 to 20 per run. Standard allows 1,000 calls a day and Pro 10,000, so a weekly rollup barely registers.

What is the difference between a monday automation and an AI agent?

An automation follows a fixed rule: when something happens, or every Monday at 9am, run the action you set. An agent reads the board, the history and the threads, works out what changed, and writes something new each time.

What does it cost, on the monday side and the Gravity side?

Gravity's Free plan is $0, Gravity's Mini Pilot plan is $15 a month and Auto Pilot is $20 a month. On monday.com, a seat costs $9 a month on the Free plan, $12 on Standard or $19 on monday's Pro plan, billed annually, from 3 seats. Weekly per-item summaries of 180 items would add about 6,235 monday AI credits a month.

Sources