SMF·MONDAY-COM
Can a prompt replace monday.com?
Project management — spreadsheet-shaped work platform
Exhibit tracking slip
Verdict
Underneath the colour, monday.com is a typed spreadsheet: every column has a behaviour, and the automations read like sentences. Rebuilding the typed-column core with a rule builder is a week, and it is a genuinely useful week because the result is the shape most small teams actually need. The gaps are the mobile apps, real-time collaboration and the template gallery — adoption machinery rather than a moat.
Exhibit A — The prompt
Received on31.07.2026Build a typed-column board platform with an automation builder that reads like English.
Stack: your choice, Postgres, Docker Compose.
Board model: a Board has Groups (row sections) and Columns. Every Column has a type from a closed set, and the type owns its behaviour — rendering, editing, sorting, filtering and aggregation: text, long text, number, status (named options with colours), date, timeline (start and end), person, checkbox, dropdown, link, file, formula, mirror (a value read from a linked item on another board), and dependency.
Summary rows: each group and the board footer show a per-column aggregate chosen by the column type — sum or average for numbers, a distribution bar for status, earliest and latest for dates, a count of ticked for checkbox.
The automation builder is the differentiator, and its constraint is that a rule must be readable as one sentence: *When [trigger], [condition], then [action].*
- Triggers: status changes to X; a date arrives; an item is created; a column changes; every period.
- Conditions: optional, one comparison on one column.
- Actions: notify a person; set a column value; move to a group; create an item on another board; archive.
Render each configured rule as that sentence in the UI. Run rules once per triggering event, log every run with the outcome, and cap the chain depth at three with a visible error rather than a silent stop — recursion is the failure mode of every rule engine.
Views over the same board: table (default), kanban grouped by any status column, calendar by a date column, and timeline by a timeline column.
Cross-board relations: a link column pointing at items on another board, and a mirror column reading one of the linked item's values. Mirrors are read-only and must never be writable through the mirror, which is the bug this feature always ships with.
Write tests for chain-depth capping, for mirror read-only enforcement, and for group aggregates with empty and null values.
Do not build a template gallery, mobile apps, or real-time collaboration.Opening prefills the prompt — press enter to run it.
Exhibit B — What you lose
- B.1 the mobile apps and push notifications
- B.2 real-time multiplayer editing on the same board
- B.3 the template gallery, which is how most teams start
- B.4 the app marketplace and its deep integrations
- B.5 AI credits and the assistants they power
Prior art
Exhibit C — Why people still pay: template ecosystem and adoption
Because monday.com is bought by people who are not going to build anything, and the templates plus the onboarding are what get a whole department using the same board within a week.
Questions
Why insist the automation reads as a sentence?
Because it is the constraint that keeps a rule builder usable by non-developers, and it is most of why monday.com gets adopted by teams who would never open a workflow tool. Once rules stop reading as sentences you have built a scripting language with a worse editor.
What is a mirror column and why is it dangerous?
It shows a value that lives on a linked item on another board. It is dangerous because making it editable creates two sources of truth for one value, and reconciling those is not a bug you fix, it is a bug you live with.
Why cap automation chain depth?
Because rule A firing rule B firing rule A is trivially easy to configure by accident, and an uncapped engine will happily send four hundred notifications before anyone notices. A visible error at depth three is better than a silent stop nobody can debug.
How does monday.com actually bill?
Per seat, in blocks, with a three-seat minimum, and the price shown by default is the annual one — the monthly rate is about 18% higher. AI features are metered in credits on top of the seat price.
Related tools
Receipt