Running Active Playbooks
Tasks tab
Every action item across the playbook in one place — list or kanban, filterable, with AI-generated tasks dropping in live as the engagement runs.
The Tasks tab is the action layer of a playbook — the running list of things that need doing in the engagement. Tasks come from a few places: tasks you create by hand, tasks captured during a session, and tasks AI-generated from session conversations or analysis steps.
Wherever they come from, they all land here, and you work them from this tab.
Two views, one task list
List
A vertical, sortable, filterable table of tasks. Best for triage, bulk scanning, and "what's on my plate this week?" reads.
Kanban
Task cards arranged in columns by status (To do / In progress / Done, etc.). Best for workflow visualisation and dragging cards to advance status.
The view toggle sits in the top-right; pick the one that suits how you think about the work. Most facilitators use list for triage and kanban for active execution.
Filters and sort
The filter bar lets you narrow what's visible:
| Filter | Useful for |
|---|---|
| Status | Hiding completed work; focusing on what's in flight or to-do. |
| Assignee | Reviewing one person's load; rebalancing across the team. |
| Due date | Surfacing what's overdue or due soon. |
| Source | Separating AI-generated tasks from human-created ones; reviewing tasks from a specific session. |
Sort options vary by view:
- List — sort by due date, status, assignee, created date, or title.
- Kanban — within each column, sort by due date or assignee.
Filters compose; the sort is single-axis.
Creating and editing a task
Click New task to open the create form. A task carries:
- Title — short and action-shaped ("Send proposal draft to client" beats "Proposal").
- Description — context, links, supporting detail.
- Assignee — who owns the task.
- Due date — when it needs to be done.
- Status — To do / In progress / Done (or whatever statuses the workspace supports).
Edit any task by clicking it to open the detail pane. Changes save as you make them — no separate save button.
AI-generated tasks and realtime updates
A nice touch on this tab: tasks that AI is generating in the background appear as placeholder skeletons that fill in as the work completes. You'll see this most often when:
- A session has just finished and Gleo is extracting action items from the discussion.
- An AI analysis step in a playbook has identified work that needs doing.
The skeletons live-update — you don't need to refresh.
Don't refresh while you see "generating" skeletons. The skeletons are deliberate placeholders showing that work is in progress. Refreshing won't speed things up; it'll just make Gleo restart the live update and you'll see the same skeletons again. Wait — it usually takes seconds, not minutes.
When the skeletons hydrate, the real task content lands in place and you can edit it like any other task. Treat AI-generated tasks as proposals — review them, refine the wording, assign owners, and confirm due dates before treating them as the team's actual to-do list.
Walkthrough
-
1
Open the Tasks tab
Inside any playbook, switch to the Tasks tab. The list view opens by default.
-
2
Try kanban
Toggle to kanban. You'll see the same tasks arranged in columns by status. Drag a card to a new column to advance its status.
-
3
Apply filters
Filter to your tasks (assignee = you), or to overdue tasks (due date in the past). Combine filters as needed.
-
4
Create a task
Click New task. Title, assignee, due date — that's the minimum useful task. Add description / detail if it helps.
-
5
Open task detail
Click any task to open its detail pane. Edit anything; changes save as you type.
Tips & pitfalls
Title tasks like actions, not topics. "Pricing discussion" is a topic; "Send pricing options to Sam by Friday" is a task. The action-shaped version self-describes its done-state, which makes triage and follow-up trivial.
Use the workspace Tasks rollup for cross-playbook scanning. The Home dashboard rolls up the top 10 tasks across your top 5 active playbooks. For the daily "what's on my plate?" view, that's faster than walking each playbook's Tasks tab in turn.
Treat AI-generated tasks as drafts. They'll usually be in roughly the right shape — but check the title, assignee, and due date before you treat them as committed work. The AI doesn't know who owns what or what the team's actual capacity looks like.