Task Analysis
Munera's AI analyses every task to extract required skills, estimate complexity, validate priority, and generate semantic embeddings for similarity search — all powered by GPT-4.
What happens when you create a task
When a task is created with a title and description, Munera's AI pipeline runs automatically in the background:
Skill extraction
GPT-4 reads the task description and identifies the technical skills required — for example React, TypeScript, PostgreSQL, or AWS Lambda. Skills are matched against your organisation's skill library.
Complexity scoring
The AI assigns a complexity score from 1–10 based on scope, ambiguity, technical depth, and estimated number of systems involved. This score influences the assignment algorithm's seniority weighting.
Priority validation
Munera cross-references the described urgency and business impact against the submitted priority level. If there's a mismatch (e.g. a described production incident filed as low), the AI flags it for manager review.
Effort estimation
Based on task description, skills, and complexity, the AI suggests an estimated effort in hours. Managers can accept or override this estimate.
Semantic embedding
A vector embedding of the task is generated using OpenAI's embeddings API. This powers similarity search — find related tasks, detect duplicates, and surface institutional knowledge from completed work.
Task lifecycle
Tasks follow a defined state machine. The current status determines what actions are available and who can perform them.
At any point a task can transition to blocked (awaiting a dependency or external factor) or cancelled.
Running AI analysis on demand
You can re-run analysis on any task at any time — useful after significantly editing a task description.
- 1Open the task
- 2Click AI Analysis in the task detail panel
- 3View updated skill suggestions, complexity score, and priority recommendation
- 4Accept or override each suggestion individually
Task priority
| Priority | Target resolution | Use case |
|---|---|---|
| critical | Same business day | Production incidents, security vulnerabilities |
| high | Within 2 business days | High-impact features, urgent bugs affecting many users |
| medium | Within 1 week | Normal feature work, non-urgent bugs |
| low | Best effort | Nice-to-haves, tech debt, documentation |
Task views
Tasks can be viewed in four layouts depending on your workflow preference:
List view
Sortable, filterable, paginated table. Best for managers reviewing the full task queue.
Kanban board
Drag-and-drop cards grouped by status. Best for sprint standups and progress tracking.
Gantt chart
Timeline view showing tasks, deadlines, and dependencies. Best for planning and scheduling.
Calendar view
Tasks plotted on a calendar by due date. Best for workload planning and deadline awareness.