Agent Performance & Conversation Quality Scoring

Velaro automatically scores every resolved conversation for quality and generates actionable coaching notes for your agents — no manual review required.

How Quality Scoring Works

After every conversation closes, Velaro's AI analyzes the full transcript and assigns:

• Quality Score (0–100) — How well the agent (or bot) handled the interaction. Based on resolution success, professionalism, empathy, and completeness.
• Sentiment — How the customer felt at the end: positive, neutral, or negative.
• Coaching Note — One specific, actionable tip to help the agent improve (or praise for a great conversation).

Scoring runs automatically in the background — it never delays the conversation closing or affects the customer experience.

Enabling Conversation Quality Scoring

Go to Settings → AI Features → Conversation Quality and toggle it on. Requires a Professional or Business plan.

Once enabled, all new conversations that resolve will be scored. Historical conversations are not retroactively scored.

Viewing the Agent Performance Report

Go to Reports → Agent Performance to see:

• Overall average quality score for your team
• Sentiment breakdown (positive / neutral / negative)
• Per-agent scores — sorted by lowest score first so you can prioritize coaching
• Recent coaching notes — expand any agent to see their most recent AI feedback
• Scored conversation table — every scored conversation with score, sentiment, and coaching note

Use the 7d / 14d / 30d / 90d filter to adjust the reporting period.

Score Interpretation

| Score | Label | What It Means |

|-------|-------|---------------|

| 90–100 | Excellent | Fully resolved, empathetic, first contact |

| 75–89 | Good | Resolved but minor tone or efficiency opportunity |

| 60–74 | Fair | Partially resolved or slow response |

| 0–59 | Needs Improvement | Not resolved, missed key points, or poor tone |

Agent AI Smart Reply

When an agent is composing a reply, they can click Suggest Reply to get an AI-drafted response grounded in your company's knowledge base. The draft appears in the reply box — the agent can edit it before sending.

Smart Reply uses the same knowledge base as your bot (product docs, help articles, custom FAQs), so suggestions are grounded in your actual content — not generic AI responses.

To enable Smart Reply, go to Settings → AI Features → Agent Smart Reply and toggle it on.

Frequently Asked Questions

Does quality scoring affect agents negatively? Scores are a coaching tool, not a disciplinary metric. They're designed to highlight growth opportunities and recognize great performance.

Can agents see their own scores? Scores are visible in the Agent Performance report, which is accessible to team managers and admins. Agent-facing score visibility is a setting you can configure.

How long does scoring take? Scoring completes within 30–60 seconds of a conversation resolving. It does not delay the resolution itself.

What about bot-only conversations? Bot conversations that never involve a human agent are still scored — the quality score reflects the bot's ability to resolve the issue, not an individual agent.

Can I adjust the scoring criteria? The scoring rubric is standardized across all accounts. Custom rubrics are on the roadmap for Enterprise plans.