SLA INTELLIGENCE CENTER
"We need to see the real picture — not just the numbers"
Here's the thing about SLA dashboards — they only tell you what happened, not why it happened. Your 95% compliance looks great on paper, but dig a little deeper and you'll find clock manipulation, workload imbalances, process bottlenecks, and assignment gaps hiding behind those numbers.
Our AI doesn't just track SLAs — it understands them. It detects when someone's gaming the system, spots overloaded engineers before they burn out, identifies which categories always breach, and recommends exactly what to fix.
And here's what makes it special: Every detection comes with a clear explanation. You'll know exactly why the AI flagged something — the data points, the patterns, the reasoning. No black boxes, no guesswork.
Fraud Detection
Workload Balance
Bottleneck Analysis
Process Optimization
Live: AI Detection Feed
🎫 TKT-4521 | P2 | Yardi System Login Failure
Property: DAMAC Hills - Golf Vista Tower
Requester: Leasing Dept | App: Property Sales Portal
Status: "Waiting for User" at 14:47
SLA Breach Due: 15:00 (13 min remaining)
⚠️ No outbound email to requester found
⚠️ Generic comment: "Waiting for user"
🤖 AI Verdict: 87% Suspicious
📋 Action: Flagged - Escalated to Fatima Hassan (Sr. Applications)
Powered by SCIKIQ GenAI
AI Detection Categories 5 Active Monitors
All Systems Healthy
SLA Gaming & Fraud
Clock manipulation, fake status changes, hidden breaches
3 High Risk
Workload Imbalance
Overloaded engineers, uneven distribution, burnout risk
2 Engineers at risk
Process Bottlenecks
Categories with chronic delays, approval backlogs
VPN avg 4.2h delay
Assignment Gaps
Unassigned tickets, skill mismatches, routing errors
7 pending >2hrs
Breach Predictions
ML-predicted breaches in next 4 hours
94% accuracy
How Our AI Detects Issues — Full Explainability Transparent AI
SLA Fraud Detection Algorithm
1
Communication Verification
35% weightWhen status changes to "Waiting for User/Approver", we check:
- Was an email sent to the user within ±2 hours of status change?
- Does the email contain ticket-specific content (not a template)?
- Is there a valid recipient email address?
IF status_change == "Waiting" AND no_email_found THEN score += 35%
2
Near-Breach Timing Analysis
30% weightSuspicious timing patterns:
- Status paused within 2 hours of SLA breach deadline
- Multiple pauses on same ticket approaching different SLA milestones
- Pattern of pauses at end of shift
IF time_to_breach < 2h AND pause_triggered THEN score += 30%
3
Comment Quality Analysis
20% weightGPT-4 analyzes comment quality:
- Generic phrases: "Waiting for response", "Pending user reply"
- Copy-paste from templates without customization
- No specific details about what was asked
IF GPT4.isGeneric(comment) > 0.7 THEN score += 20%
4
Engineer History Pattern
15% weightHistorical behavior analysis:
- Number of previous flagged tickets (30 day window)
- Pause frequency compared to team average
- SLA save rate (breaches avoided by pauses)
IF previous_flags > 2 THEN score += 15%
Live Detection: DAMAC Leasing Ticket
TKT-4521 | Yardi Login Issue
87% Suspicious
DAMAC Hills - Golf Vista | Leasing Dept | Assigned: Mohammed Khan
No email sent to leasing agent
+35%
Paused 13 min before P2 breach
+30%
Generic: "Waiting for user response"
+15%
Engineer: 2 prior flags this month
+7%
AI Reasoning: High confidence of SLA gaming. Leasing agent at Golf Vista reported Yardi login failure affecting property viewings. Status changed to "Waiting" 13 min before breach with no email contact. Pattern matches 2 prior incidents from same engineer affecting Property Sales Portal operations.
Detection Thresholds
Auto-Flag Threshold
Tickets above this score are flagged
40%
High Risk Threshold
Requires immediate manager review
70%
Email Window
Time window to check for emails
±2 hours
Real-Time SLA Health Command Center LIVE
Updated 30s ago
SLA Compliance: Reported vs True
95%
Reported
(What dashboard shows)
82%
True (AI-Adjusted)
(After fraud removal)
13% gap = Hidden SLA breaches
True SLA by Priority
Breach Predictions (Next 4hrs) AI
TKT-4892 - Payment Gateway
P1 - Property Sales Portal | Downtown Dubai
94%
in 47 min
TKT-4901 - Yardi Timeout
P2 - Leasing Team | DAMAC Hills
76%
in 1.5 hrs
TKT-4887 - CRM Sync
P2 - Customer Relations | Business Bay
58%
in 2.3 hrs
SLA Fraud Impact Calculator Business Value
$287K
Hidden SLA Penalty Exposure
Per Year (if undetected)
$341K
Saved by Early Detection
Penalties avoided this year
Fraud Impact Breakdown
23
Hidden P1 Breaches
156h
Suspicious Pause Time
8
Engineers Flagged
Contract Protection
Accurate SLA data for vendor negotiations
+12% leverage
GenAI Decision Explainability Transparent AI
TKT-4521 | Yardi System Login
87% Suspicious
DAMAC Hills - Golf Vista | Leasing Dept | Property Sales Portal
No Outbound Email
Status changed to "Waiting" but no email sent to Leasing agent
+35% to suspicion score
Near-Breach Timing
P2 SLA paused with only 13 minutes remaining - Property Sales impacted
+30% to suspicion score
Generic Comment
"Awaiting customer response" - matches placeholder pattern
+15% to suspicion score
Engineer: Mohammed Khan (Service Desk)
2 previous flags on Property Management tickets this month
+7% pattern bonus
AI Recommendation
Escalate to Service Desk Manager. This Yardi ticket affecting DAMAC Hills leasing operations shows SLA clock manipulation. Recommend: 1) Correct SLA timer (+13 min), 2) Coaching session with Mohammed Khan, 3) Alert Fatima Hassan (Sr. Applications) to verify Yardi system status.
Enterprise Compliance & Audit Trail SOC2 Ready
2,847
Tickets Audited (30 days)
47
Fraud Cases Detected
98.3%
Audit Accuracy Rate
24/7
Continuous Monitoring
DAMAC Properties - Recent Audit Events
FLAG
TKT-4521 (Yardi Login - DAMAC Hills) flagged for SLA manipulation
2 min ago
ALERT
Mohammed Khan (Service Desk) - 3 flags threshold | Manager notified
15 min ago
CLEAR
TKT-4498 (Payment Gateway - Business Bay) reviewed - legitimate pause
1 hr ago
BOTTLENECK
Property Sales & Marketing SLA breach rate increased 8% this week
2 hrs ago
REPORT
DAMAC Q4 SLA compliance report generated - sent to Rashid Al-Habtoor
3 hrs ago
Agent Journey - Fraud Detection Pipeline
0 of 5 steps
1
Data Collection
Pending
Scan all tickets with "Waiting for User/Approver" status changes and retrieve
communication logs and support engineer activity data.
0
Tickets
0
Pause Events
2
Pattern Analysis
Pending
Apply fraud detection rules to identify suspicious patterns including
missing communications, near-breach pauses, and placeholder comments.
0
Rules Hit
0
No Comms
3
Risk Scoring
Pending
Calculate suspicion score based on weighted fraud indicators,
support engineer history, and ML-based confidence scoring.
0%
Avg Score
0
High Risk
4
Flag & Alert
Pending
Flag tickets with score > 40% as suspicious, generate investigation
reports, and calculate SLA accuracy impact.
0
Flagged
0h
Sus. Hours
5
AI Deep Analysis
Pending
SCIKIQ GenAI analyzes each flagged ticket to provide detailed
reasoning, evidence assessment, and actionable recommendations.
0
Insights
0
Support Engineers
Fraud Detection Algorithm Weights
No Outbound Communication
35%
Near-Breach Timer Pause
30%
Generic/Placeholder Comments
20%
No User Response Evidence
15%
Agent Activity Log
0 events
Waiting to start journey...
Real-time Metrics
-
Flagged
-
Sus. Hours
-
Hidden SLAs
-
SLA Impact
-
Flagged Tickets
-h
Suspicious Pause Time
Hidden Breaches
-%
SLA Accuracy Impact
Agent Flagged Tickets
Detection Rules
No Communication
Status changed without email to user
Near-Breach Pause
Timer paused within 2h of SLA breach
Generic Comments
Placeholder text without specifics
No User Response
Extended wait without reply evidence