Scoring with playbooks

A playbook is your scoring rubric. It's a named set of questions — each with a category and a severity — that Perceive8 answers for every captured conversation. Playbooks are how "we had a call" becomes "the call scored 82, with one critical compliance miss."

What a playbook looks like

Each question in a playbook has:

• Question — what to check, in plain language: "Did the agent read the recording disclaimer?"
• Category — how you group it: Compliance, Discovery, Objection handling, …
• Severity — how much it matters: INFO, WARNING, or CRITICAL

When a conversation is analyzed, every question gets an answer with timestamped evidence from the transcript — so a score is never a black box; you can always click through to the exact moment that produced it.

Start from a template

The fastest path is cloning a built-in template. Perceive8 ships 16, including:

• Sales Coaching — discovery, qualification, next steps
• Meeting Intelligence — decisions, action items, engagement
• Legal Depositions, Mortgage Compliance — regulated-conversation checks
• CRM Contact Update, Follow-up Email, Manager Escalation — outcome-focused workflows
• Generic Action Items — a good default for any team

1. Go to Audio → Playbooks.
2. Find a template close to your use case and click Clone & Customize.
3. Rename it and adjust the questions to your wording.

Create one with Aria

You can also describe your rubric in plain language:

1. Go to Audio → Playbooks and click + Create with Aria.
2. Tell Aria what a good conversation looks like for your team — "score my support calls on greeting, empathy, resolution, and whether a follow-up was offered, and flag any refund promise as critical."
3. Review the questions Aria proposes and refine them in the chat.

Activate a playbook

Only one playbook is active at a time per workspace. The active playbook:

• Scores every new upload automatically
• Scores auto-joined meetings (you can override it per meeting in the auto-join section)

Click Activate on a playbook to make it the default; the banner at the top of the page always shows which one is live, and you can deactivate it there.

What scoring produces

Scoring a conversation generates a report:

• Score (0–100) with sub-scores for compliance, talk ratio, and alert load
• Overall assessment — All clear, Minor concerns, Significant concerns, or Critical issues
• Per-question findings — each answer with severity and timestamped transcript evidence
• Executive summary and recommendations

You'll find reports in the Scorecard, Actions, and Playbooks panels of any analysis, and you can generate an Aria Report on demand from the Playbooks panel.

Tips

• Start narrow. Five to ten well-worded questions beat forty vague ones.
• Reserve CRITICAL for things that genuinely need escalation — it drives alerts during live streams.
• Iterate after a week. Read a few reports, note questions that never add signal, and refine with Aria.

Next steps

• Understanding your analysis
• Use cases and skills API — manage playbooks programmatically
• Aria assistant