Prodedify
Kiln AI

Issue evaluation

AI-powered issue quality scoring and improvement suggestions.

Kiln can automatically evaluate issues for quality and completeness, helping your team maintain a high-quality backlog.

Clarity scoring

When enabled, Kiln analyzes each issue and assigns a clarity score (0–100).

What's measured

The score considers several factors:

  • Is the problem or goal clearly described?
  • Are acceptance criteria defined?
  • Is there enough context for someone to start working?
  • Are edge cases or constraints mentioned?

How scores appear

Issues with low clarity scores are flagged with a badge in list and board views, making it easy to spot issues that need improvement before they enter a sprint.

Improvement suggestions

Alongside the score, Kiln provides specific, actionable suggestions.

Types of suggestions

  • Add a clearer description of the problem
  • Define acceptance criteria or success metrics
  • Break the issue into smaller, more focused pieces
  • Add technical context or constraints
  • Clarify scope or out-of-scope items

Evaluation snapshots

Each evaluation is saved as a snapshot, creating a history of how issue quality changes over time.

Tracking improvement

Compare snapshots to see if your team's issue writing is improving. Re-trigger evaluation after updating an issue to get a fresh score.

Snapshot history

View the full evaluation history for any issue from the issue detail panel.

Configuration

Issue evaluation is configured per project.

Enabling evaluation

Go to Project Settings → AI to enable or disable automatic evaluation.

Model selection

Choose which AI model to use for evaluation. Different models offer different tradeoffs between speed and quality.

Build readiness

Beyond clarity, Kiln also assesses whether an issue is ready to build.

What "ready to build" means

An issue is ready to build when it has enough detail for the AI agent to start working on it autonomously — clear requirements, defined scope, and sufficient technical context.

Starting a build

When an issue is marked as build-ready, you can click Develop to launch Kiln AI with the issue pre-loaded as context.

On this page