How to Use the Voice AI Prompt Evaluator in Sparkforge

The Voice AI Prompt Evaluator sits inside Sparkforge's Voice AI module and gives you actionable feedback on agent prompts. You can apply suggestions, dismiss ones that do not apply, and re-evaluate as you tune the prompt.

What it does

• Real-time checks with specific, actionable suggestions.
• Resolve to apply a recommendation or waive to dismiss it.
• One-click re-evaluation after edits.
• Filter to high-priority items only when you want to focus on critical fixes.

How to run an evaluation

1. Log into your sub-account.
2. In the left navigation, click AI Agents.
3. Click the Voice AI tab.
4. Decide whether to edit an existing agent or create a new one.
5. On the agent's row, click the three dots next to Last updated.
6. Click Edit to open the agent editor.
7. Open the Agent Goals section.
8. If the agent is in Basic mode, click Switch to Advanced mode.
9. In the Prompt area, click Evaluate.

Working through results

The Evaluation results panel opens on the right with suggested changes and findings.

• Click any item to expand its details and recommendation.
• Click Resolve to accept and apply the recommendation.
• Click Waive to dismiss a suggestion that does not fit your scenario.
• Toggle Show critical only to filter to high-priority items.

Re-run Evaluate as many times as you want. The tool is built for iterative refinement.

How to use it well

Treat each evaluation as one round of testing, not a final answer. Apply the changes that actually fit your agent's job, waive the ones that do not, then re-evaluate. The tightest prompts usually take three or four passes, with critical-only filtering to stay focused on real issues.