Configurable boards and docs with reporting, rule automations, Sidekick, AI blocks, multi-step workflows, custom agents, and Vibe apps.
monday.com
Explore features, practical uses and pricing below.
monday.com combines a configurable work platform with AI assistance, automation, agents, and custom app creation. Teams use boards to track work, documents to explain it, and views and dashboards to review progress. AI features can then summarize information, classify requests, help maintain boards, or perform tasks using permitted context and tools.
The platform is particularly useful when work needs a shared structure: owners, dates, status, dependencies, and a clear handoff between teams. A marketing operations group can manage campaign requests; a project team can track deliverables; an operations team can organize recurring work. The AI capabilities build on that structure, so a well-defined board remains important even when you use an assistant or agent.
The AI work platform guide describes workspaces containing related boards, dashboards, and docs. A board uses items for the things you track, groups to organize them, and columns for their details. Templates provide starting structures that you can adapt to your own process.
A campaign request board might contain the request title, requester, owner, requested date, brief, approval status, and delivery link. A project board might use milestones and dependencies instead. Decide what an item represents before adding columns: a campaign, an asset, and a single task are different units of work.
The board basics explain the underlying elements. Keep labels meaningful to the team. If “Done” means delivered on one board and approved on another, a cross-team summary can conceal a real difference. Written definitions help people and AI-powered workflows interpret the same information consistently.
Views let a team examine the same board through a different lens. A daily execution view and a planning view do not need to show the same fields. Use the format that helps answer the question at hand rather than treating every available view as necessary.
Dependencies describe relationships between tasks and how their dates interact. Configuration and rollout affect the available dependency options. The documentation explains that dependencies shift existing dates rather than filling every blank date automatically. If a chain looks incomplete, check the dates and duration setup before assuming the dependency failed.
Dashboards aggregate data from connected boards through widgets. A widget can only display the relevant information that exists in those boards, and access and plan limits affect connections. A project overview field outside the board-column data model may need to be brought into a column before it can appear in the desired report.
For a leadership dashboard, agree on the reporting question first: blocked projects, overdue deliverables, or workload. Then check the contributing boards and fields. A clean chart is less useful if half the projects lack the column it uses.
Workdocs provide collaborative documents with comments, content elements, and connections to board work. They can be linked through a Docs column and can turn text into board items. Sharing and edit permissions determine who can see or change the document; supported export options provide another way to share a finished version.
A delivery board needs concise operational fields, but a project brief needs room for reasoning. Use a workdoc for objectives, assumptions, and acceptance criteria, then link it to the relevant item. That keeps the reason for the work available without turning every board cell into a long document.
Sidekick can help summarize or develop document content. Review whether the summary preserves decisions and exceptions. A short project description that omits a constraint can make the next task look simpler than it actually is.
monday sidekick is the conversational assistant. It can work at account, board, and item levels, use supplied and permitted context, research information, create content, and perform supported board operations. Connected skills extend its work to services such as email, messaging, and calendars. The account configuration determines the available scope.
Use a specific request tied to the board you mean. For example, ask which launch items have a due date this week and an unresolved approval, then request a draft update grouped by owner. Distinguish the analysis from the action: finding blocked work is different from changing dates or sending the update.
Sidekick also provides personalization, custom guidelines, and editable memory. Durable preferences can help with recurring work, but temporary project details can become outdated. If a recommendation uses an old assumption, check the relevant remembered information and restate the current constraint before asking for another plan.
AI blocks support actions within columns, automations, and workflows. Documented categories include assigning labels, summarizing text, improving writing, translation, sentiment, information extraction, document generation, and custom instructions. Prompt previews help you inspect a result before wider use.
An intake board could summarize a long request into a short field and classify it into an existing category. Define the categories and what should happen when the request does not fit. Without that instruction, a classification can look neat while pushing an unusual request into an inappropriate queue.
Keep extraction tied to the source. If a brief contains a proposed date and a confirmed date, instruct the block which one to use and how to flag ambiguity. A sentiment label can help prioritize review, but it should not replace reading an important complaint. The useful result is a field that assists the next decision, not merely another generated column.
Board automations apply triggers, conditions, and actions to routine work. They can update items, notify people, or move work when the selected condition occurs. Templates and custom recipes provide setup routes, and an automation can be toggled off without deleting its configuration.
Use a straightforward rule when the decision is already known. If an approved request always moves to a production group, a rule is easier to inspect than asking an agent to decide where it belongs. AI is more relevant when the input requires interpretation, such as turning an unstructured brief into a draft summary.
Review overlapping rules together. A status change that moves an item can also trigger another action on the destination board. Test the complete sequence and check that it does not produce duplicate notifications or conflicting updates. Give the automation a name that explains the business purpose so its owner can find it later.
The AI workflow builder supports a visual process made from triggers, actions, conditions, waits, and delays. You can describe a process to generate a starting workflow or build it manually. It supports multi-step and cross-board work, with AI-powered blocks available for supported tasks. Product and plan requirements differ.
Review every generated block before publishing. Check the triggering event, target board, mapped fields, and branches. When a later step refers to an earlier item, confirm that it uses the intended item rather than a similarly named record. The Run now option and run history help inspect the sequence and locate a failed step.
An operations workflow might receive a request, summarize its brief, check its priority, and route it into a review queue. Include a branch for incomplete information. Keep an explicit review status between a generated recommendation and an action that represents a final decision.
The agent configuration guide separates core instructions, knowledge and access, tools, skills, and model choices. Board access can distinguish reading from editing. Uploaded material can supply additional context, while tool guardrails can restrict actions to particular channels or recipients.
Jobs divide the agent's work into focused tasks with their own triggers and instructions. They can include recurring schedules. That is useful when one agent needs to produce a weekly brief and respond to a direct assignment, because those two tasks can have different inputs and endpoints.
Describe both the intended result and the limits. An agent maintaining a request queue might summarize items and propose owners while leaving priority decisions to the team. If you later allow updates, enable the required access deliberately and test that exact operation. Granting a broad tool does not replace instructions about which records should change.
The agent management guide describes assignment-based use, activity records, memory, sharing, and controls. Activity shows run status, tools used, and credit consumption. Failed or partial runs provide details that can help identify where a task stopped. The documented availability is scoped to the AI platform, with broader product availability evolving.
Review a run's actual changes, not just its success label. A run can finish after producing a draft when the operator expected a completed handoff. Agree on that endpoint in the job instructions and check a sample of outputs after changes to sources, board structure, or connected tools.
Agent memory can carry context between runs. Keep enduring process rules separate from temporary facts such as this week's campaign deadline. When sharing an agent, confirm who can use it and how the configured access and usage attribution apply.
monday vibe turns prompts into custom business apps within the monday environment. It can use selected board data and be refined through conversation. You might build a specialized request view, an internal resource display, or a dashboard that presents a team's board information in a different form.
Specify what the app should display and what users may change. A read-only request overview is a different build from a form that updates an item's status. Test the key fields and the behavior for an empty board or an item with missing information before sharing the app.
The Vibe FAQ describes product-specific boundaries: apps currently live inside monday.com, run while actively viewed, and do not provide the same background-job behavior as scheduled agents. Its documented external API and export limitations also matter if you are planning an independent application. A board change made by an app can trigger a board automation, which is different from the app directly running an automation.
Start with a board where each item represents one requested campaign asset. Require a brief, intended audience, requested date, and requester. Add separate fields for a generated summary, proposed category, assigned owner, and human approval. Keep the approval field distinct from the AI's recommendation.
Use an AI block to summarize the brief and propose a category from your team's defined labels. Have an operations reviewer confirm the category and owner. A board automation can then notify the owner when the request reaches the agreed review status. If routing crosses several boards, use a workflow and inspect the field mappings.
Sidekick can help draft a weekly queue summary. A custom agent can handle a recurring brief once its sources, schedule, and endpoint are configured. If reviewers need a more specialized interface, evaluate a Vibe app over the same board. Add those layers because they solve a specific problem in the process, rather than replacing a working board all at once.
Platform permissions determine access across account, workspace, board, and other levels. Main, private, and shareable boards serve different audiences, with plan requirements for some types. A workspace cannot grant broader access than the account policy allows.
AI governance adds feature access and usage controls. Enterprise supports more granular role and selected workspace controls. Administrators can inspect consumption and manage limits. Changing a permission is not necessarily a stop command for an already running block or agent; use the relevant operating control when a process needs to be paused.
For a shared workflow, identify the board owner, connection owner, and person responsible for investigating a failed run. Those responsibilities can belong to different people. A departing colleague's connection or a changed board permission can affect a process that previously worked.
The pricing page offers a limited free tier and paid plans, with AI access beginning in the paid offering rather than being identical across all accounts. Seat counts, billing term, product, and selected capabilities affect the purchase. CRM, dev, service, and the AI work platform have distinct commercial and feature scopes.
AI credits measure supported AI usage in different ways. Blocks, Sidekick, workflows, and agents do not all consume an identical flat amount. Complexity and the work performed can affect usage, and administrators can inspect feature consumption and manage allowances.
Vibe access and publishing add another distinction: building, AI-powered app activity, and publishing have their own requirements. Publishing requires a Vibe app package. Confirm the current allowance and publication arrangement before committing a shared app to a team process.
No. Boards, docs, views, and rule-based automations provide the underlying work structure. Add AI where interpretation, drafting, or a defined delegated task helps your process.
No. Vibe provides a custom application interface; agents can have configured tasks and recurring jobs. Check the supported operating behavior before choosing either for work that must continue in the background.
No. Product, paid plan, administrative enablement, credits, and publishing packages determine access. Review the specific feature you intend to operate.