Docs

AI Smart Assist

Leverage intelligent AI assistance to polish feedback tone, structure actionable bug reports, and synthesize threads.

What is AI Smart Assist?

AI Smart Assist helps teams communicate feedback more constructively and transform informal thoughts into structured, developer-ready bug reports. Instead of spending minutes manually writing step-by-step reproduction instructions or softening constructive feedback, Smart Assist does it in one click.

Polishing Comment Tone

Feedback between cross-functional teams is most effective when it is objective, polite, and actionable.

  1. Type your rough feedback or bullet points into the comment composer.
  2. Click the AI Sparkle icon below the text field.
  3. Choose a tone preset:
  • Constructive & Polite: Softens harsh phrasing into thoughtful design and product recommendations.
  • Concise & Direct: Distills wordy feedback into brief, high-impact bullet points.
  • Client-Friendly: Formats feedback into polished client presentation language.

Generating Structured Bug Reports

Turn casual remarks like "the dropdown closes whenever I try to select an option" into professional engineering tickets:

  1. Click Make Actionable in the Smart Assist menu.
  2. Smart Assist analyzes the target element's DOM context, viewport dimensions, and user text.
  3. It generates a structured markdown report including:
  • Observed Behavior: Clear summary of the issue.
  • Expected Behavior: What should occur according to design patterns.
  • Technical Details: Captured DOM selector and current viewport.
  • Suggested Remediation: Potential CSS or JavaScript fix.

Thread Summaries

When lengthy comment threads accumulate multiple replies between engineers, QA, and clients, use Smart Assist to synthesize the conversation:

  • Open the thread overflow menu (···).
  • Select Summarize Thread.
  • An AI-generated badge appears with the key decisions made, outstanding questions, and the agreed action item.

Credit Consumption

AI Smart Assist operations utilize advanced language models:

  • Token-Based Billing: Credits are deducted dynamically based on the exact input and output tokens processed for your rewrite or summary.
  • Preview & Estimation: Smart Assist shows a real-time token and credit estimate before executing complex summaries.
  • Deterministic review features (such as manual comments, pinning, and status changes) are completely free and consume 0 credits.