FreightConnect

Agents / Agent Health

Monitoring Agents & Logs

Monitor agent performance, review decision logs, and optimize behavior.

Agent Health Dashboard

Go to Analytics → Agent Health to see:

Quoting Agent

  • Avg quotes per load: (target: 8+) higher = better market competition
  • Quote response time: (target: < 90 sec) faster = more real-time
  • Acceptance rate: (target: 75%+) % of quotes you accept
  • Margin analysis: avg margin per load, margin distribution

Drill down by:

  • Route (which lanes get better quotes?)
  • Carrier (which carriers quote most competitively?)
  • Time of day (quotes better at certain hours?)

Carrier Vetting Agent

  • Carriers approved: total in your network
  • Carriers escalated: flagged for manual review (usually new carriers or violations)
  • Rejection rate: % blocked due to failed compliance
  • Re-vetting accuracy: did carriers flagged for issues actually have problems?

Tracking Agent

  • Exception detection accuracy: (target: > 85%) did we correctly predict late deliveries?
  • False positive rate: (target: < 10%) alerts that didn't materialize
  • Shipper satisfaction: (target: > 90%) did shippers find tracking helpful?
  • Average prediction time: how far in advance do we flag exceptions?

Customer Service Agent

  • Resolution rate: (target: > 75%) % of inquiries handled without escalation
  • Escalation rate: (target: < 25%) % escalated to humans
  • Response time: avg time to first response (target: < 5 min)
  • Satisfaction score: shipper rating of agent helpfulness (1-5 stars)
  • Containment: % fully resolved without follow-up

Agent Logs

Go to Analytics → Agent Logs to see detailed decision records:

Log Entry Format

Timestamp: 2026-04-13T14:30:00Z
Agent: Quoting
Event: load.posted
Load ID: load_xyz123
Input: {
  "origin": "Los Angeles, CA",
  "destination": "New York, NY",
  "weight_lbs": 25000,
  "equipment_type": "dry_van",
  "asking_rate": 2.15
}
Decision: Query 42 carriers (filtered by FSC >= 70, insurance verified)
Action: Sent quote requests
Result: 15 quotes received in 87 seconds
Top 3: [
  {"carrier": "ABC Trucking", "rate": 2.05, "fsc": 82},
  {"carrier": "XYZ Logistics", "rate": 2.12, "fsc": 85},
  {"carrier": "Fast Freight", "rate": 2.15, "fsc": 72}
]
Your action: Accepted XYZ Logistics at $2.12/mile
Outcome: ACCEPTED ✓

Log Features

  • Filter: by agent, event type, date range, load ID
  • Search: find specific decisions or outcomes
  • Export: download logs as CSV for analysis
  • Feedback: mark an entry as "Correct," "Incorrect," or "Neutral"

Mark entries as "Incorrect" to train the agent. It learns from feedback.

Searchable Log Events

  • load.posted — load was posted
  • quote.received — carrier quote came in
  • quote.ranked — quotes were ranked for display
  • carrier.vetting_complete — vetting finished
  • carrier.escalated — flagged for manual review
  • shipment.exception_detected — tracking exception found
  • shipment.alert_sent — you were notified
  • customer_inquiry — shipper sent message
  • response_generated — agent generated response
  • response_escalated — escalated to human

Performance Trends

Go to Analytics → Agent Trends to see:

  • Over time: how agent performance changes week-to-week
  • By metric: quote volume, escalation rate, resolution rate
  • Comparisons: this month vs. last month

Use trends to identify:

  • Improving agents (keep current config)
  • Declining agents (adjust thresholds, check for external factors)
  • Seasonal patterns (busy times = different behavior)

Intervention Audit Trail

Every time you approve/reject an agent recommendation, it's logged:

Timestamp: 2026-04-13T14:45:00Z
Agent: Carrier Vetting
Recommended action: BLOCK carrier "New Carrier Inc" (fraud risk = 72)
Your decision: OVERRIDE — Approve anyway
Reason: "Known carrier, false flag on ownership screening"
Outcome: Carrier approved; no issues since then

This helps you understand your override patterns and when the agent is right vs. wrong.

Custom Dashboards

Create custom reports:

  1. Go to Analytics → Custom Report
  2. Select metrics: quoting volume, FSC distribution, exception accuracy
  3. Select dimensions: by route, carrier, time of day
  4. Set date range
  5. Download as PDF or PNG

Share reports with your team or board.

Alert Configuration

Set up alerts for agent issues:

  1. Go to Settings → Alerts
  2. Enable alerts for:
    • Quote volume drops below X
    • Escalation rate exceeds threshold
    • Agent error (if agent fails)
    • Shipper complaint (if satisfaction drops)

Get notified via email or SMS when issues arise.

Agent Restart & Reset

If an agent misbehaves:

  1. Go to Settings → Agents → select agent
  2. Click Pause Agent (stops it immediately)
  3. Fix the configuration
  4. Click Resume Agent

Or Reset to Defaults to revert all customizations.

Data Retention & Export

  • Logs retained: 90 days (live dashboard)
  • Archived logs: 2 years (accessible but slower)
  • Export: download logs for long-term analysis or compliance

Logs are kept for audit and learning. You can request deletion if needed.

Troubleshooting

Agent performance degraded: Check the logs for clues. Is it receiving bad data? Are thresholds off? Review recent configuration changes.

Too many false alerts: Lower the sensitivity (increase delay threshold, route deviation radius, etc.)

Logs not showing: Make sure the agent is enabled. Paused agents don't generate logs.

Can't find a specific log entry: Use the search filter. Logs are indexed by load ID, carrier name, agent, and event type.

Next Steps

Last updated: April 2026