MazBot — Team Performance & QA Guide

1. What this feature does

Two parts

2. Who can use it

If you do not see Team in the sidebar, ask your account owner to grant the manage team permission.

3. Where to find it

In the client panel sidebar, open Team:

Additional pages: QA Scores (per-conversation list); View Activity per agent (detailed log + KPIs).

4. Quick start (5 minutes)

Step 1 — Open Team Performance

1. Go to Team → Team Performance & KPIs
2. Set Date from / Date to (default: last 30 days)
3. Click Apply filters

You should immediately see response time columns for each agent (no extra setup).

Step 2 — Enable QA (optional but recommended)

1. Go to Team → QA Settings
2. Turn on Enable QA Scoring
3. Review defaults: greeting phrases, forbidden words, sale closed tags (e.g. sale, بيع)
4. Click Save

Step 3 — Run first scoring batch

1. On QA Settings, click Score Recent Conversations
2. Go to QA Scores to review results
3. Return to Team Performance — QA columns populate after scoring

5. Response time KPIs

These metrics use human agent messages only (message_by = agent). Bot and AI replies are excluded.

Tips

• Median is often more fair than average for coaching.
• Use date filters for weekly/monthly reviews.
• Compare agents on Team Performance, then drill into View Activity.

Current limitations

• Counts all hours (not only business hours)
• Counts replies from any assigned agent in the thread
• Does not separate WhatsApp / Telegram / Messenger yet in the KPI view

Business-hours SLA and channel-specific reporting can be customized after you confirm requirements with MazBot support.

6. QA scoring — setup

Path: Team → QA Settings

Greeting phrases

Comma- or line-separated. The agent's first reply must contain at least one phrase.

مرحبا, أهلا, السلام, hello, hi, welcome

Forbidden words

Words agents must not use. If any appear in any agent message, the forbidden-words score is lost.

Sale closed tags

If the contact has a matching tag (case-insensitive), the conversation counts as sale closed.

sale, sold, closed, won, بيع, تم البيع, مبيع

Tag a contact: Inbox → contact sidebar → add tag after a successful sale.

Sale closed contact attribute (optional)

1. Create an attribute in Contacts → Contact Attributes (e.g. Order Status)
2. In QA Settings, set Sale Closed Contact Attribute
3. Set Accepted Attribute Values (e.g. yes, نعم, تم, مباع)

Scoring weights

Maximum score = 100 (sum of earned weights).

7. QA scoring — how scores are calculated

1. Loads all non-campaign messages for the contact
2. Filters to messages sent by the assigned agent
3. Checks greeting, forbidden words, and sale status
4. Saves one row in QA Scores (updated if you score again)

Per-conversation result

Per-agent rollups

8. Reading the dashboards

Team Performance & KPIs

One row per agent. Suggested review order: median response time → QA avg score → sale conversion rate → forbidden violations.

Individual agent — View Activity

Open from team list → ⋮ → View Activity. Shows activity log plus Response Time & QA KPIs for the filtered date range.

QA Scores

Table of every scored conversation: contact, agent, score / 100, greeting, forbidden count, sale closed, scored at. Filter by agent and date range.

9. Sale conversion tracking

MazBot does not automatically know a sale happened unless you mark it.

Recommended workflow

1. Agree on tags, contact attribute, or future integration
2. Train agents to mark deals when they close
3. Run QA scoring (manual or nightly)
4. Review Sale conversion rate on Team Performance

10. Running scoring (manual & automatic)

Manual

Team → QA Settings or QA Scores → Score Recent Conversations

• Scores up to 200 most recent conversations
• Safe to run multiple times — existing scores are updated

Automatic (server)

Scheduled daily at 02:45 (requires cron + php artisan migrate on deploy):

php artisan qa:score-conversations

# One client only:
php artisan qa:score-conversations --client=123 --limit=200

Only clients with QA Scoring enabled are processed. Ensure cron runs php artisan schedule:run every minute.

11. Exporting reports

Team Performance → Export CSV includes per agent: conversations handled, avg/median/first response, QA avg score, sale %, greeting %, forbidden violations. Apply date filters before export.

12. Best practices for managers

Weekly routine

1. Open Team Performance (last 7 days)
2. Export CSV
3. Review QA Scores for low scores or violations
4. Coach agents using inbox examples

13. FAQ & troubleshooting

Response times but empty QA columns

• Enable QA Scoring in QA Settings
• Click Score Recent Conversations
• Ensure agents sent messages (message_by = agent)

QA score is 0 or very low

Does this use AI credits?

No. Current QA is rule-based (phrases, words, tags).

14. Settings to confirm with your team

15. ملخص عربي

أين أجدها؟ القائمة الجانبية → فريق العمل (Team) → أداء الفريق والمؤشرات / إعدادات الجودة

1. افتح أداء الفريق لرؤية سرعة الرد
2. فعّل تقييم الجودة من الإعدادات
3. أضف الكلمات الممنوعة ووسوم البيع
4. اضغط تقييم المحادثات الحديثة
5. راجع درجات الجودة ونسبة إغلاق البيع