AI Workflow Dashboard: Track Progress, Blockers, and Review

An AI workflow dashboard is a shared view of the current state of AI-assisted work: what is active, what is blocked, what is waiting for review, and who is responsible for each item. Without it, the team relies on memory and direct messages to know what stage each task is in. Tools like fenbs provide that shared task state through a kanban board, attributed history, and a record of where each task was left.

What an AI workflow dashboard tracks

The core job of a dashboard is to answer four questions without a meeting: what is being worked on now, what has stopped and why, what needs a person to check before it moves forward, and what is waiting on someone else.

• Active work: tasks with a current owner, an AI or person working on them, and a clear next action.
• Blocked steps: tasks that have stopped because of a missing input, an unanswered decision, or a tool that is not yet set up.
• Pending reviews: outputs from an AI step that a person must check before the workflow continues.
• Waiting items: tasks paused while an external person responds or a process outside the team completes.

A dashboard that covers only active work misses the two most important categories: the blocked steps that slow everything else, and the pending reviews that are easy to forget. Both categories are where AI workflows quietly stop rather than visibly fail.

A simple AI workflow dashboard layout

This layout works whether the dashboard is a kanban board, a shared spreadsheet, or a dedicated project tool. The columns matter less than the habit: every task has a clear lane and a stated reason when it moves from Active to Blocked or Pending review.

Tracking active work in an AI workflow

Active work should show enough context for anyone to understand what is happening without asking. The task description, the current owner, the last update, and the next action together tell the story. If those four things are not present, the dashboard is a list, not a status.

For AI-assisted steps, note which model or tool is working on the task and what it is doing. "Claude drafting the summary from the Q3 report" is more useful than "in progress." When a person picks it up for review, they know exactly what to expect.

The jira-dashboard guide covers how to structure boards for software teams. The same principle applies here: the board should reflect the real state of the work, not an optimistic summary of it. A task marked "in progress" that has not moved in three days is blocked, and the board should say so.

Flagging blocked steps clearly

A blocked step needs a clear reason and a named owner. "Blocked" without explanation stops the rest of the team from knowing whether they can help or whether to wait. Write what is missing, who needs to provide it, and when you expect it.

Blocks that stay on the board for more than a day need escalation. An AI workflow that depends on a model, a service, or a file that is not available needs a visible flag rather than a silent stall. The ai-agent-approval-workflow guide explains how to build the approval step so that a block surfaces to the right person quickly.

Managing pending reviews

Pending review is the most common place where AI work gets lost. An AI produces output, nobody checks it, and it either sits unused or moves forward without verification. A dashboard that makes pending review visible treats it as required work rather than optional tidying.

Assign a reviewer at the point the task enters pending review. A task with no named reviewer stays in that lane indefinitely. The reviewer checks the output against the criteria defined when the task was created, not against a general impression of quality. Record the review decision so the team can see whether the output was approved, revised, or rejected.

Reading the dashboard with your team

Review the dashboard at the start of a session, not at the end. A brief read of blocked and pending items before the session begins means you can address them in the same session rather than discover them at the end when there is no time left.

Keep the dashboard honest. A task that is "in progress" but has not moved in three days is blocked, not active. A review that has been "pending" for a week has not been assigned. The dashboard only helps the team if it reflects what is actually happening. For advertising workflows and other high-output processes, a stale dashboard creates more confusion than having no dashboard at all.

How to do this with fenbs

fenbs organizes work on a kanban board with four lanes: To Do, Next Up, In Progress, and Completed. Tasks record their plan, progress, and history. AI apps connect over MCP and update the board as they work, so the dashboard stays current without manual updates after every step.

1. Create a free account. Set up a project for the workflow you want to track.
2. Add each current task with a description, an owner, and the next action. Move blocked tasks to a clear status and write the reason in the task.
3. Use the In Progress lane for active work. When a step completes and needs review, add a note naming the reviewer and what they are checking.
4. Connect your AI app from Integrations. Ask it to read the project context and update tasks as it completes steps.
5. At the start of each session, review the board for blocked and pending items. Check History for recent changes before acting on a task.