The hypothesis
Field service operations track first-time-fix rate religiously. It's the metric that determines technician utilisation, customer satisfaction and profit margins. Yet most operations directors stare at dashboards showing FTFR stuck at 70-75% for months without understanding why.
The hypothesis: an agent could analyse dispatch patterns, technician performance and job characteristics to surface the specific factors causing repeat visits. Instead of guessing whether it's parts availability, skill mismatches or job complexity, the agent would quantify each factor's contribution to FTFR stagnation.
Manual root cause analysis takes operations managers 8-12 hours per week - pulling reports from the field service platform, cross-referencing technician certifications, analysing parts usage patterns. By the time insights emerge, the patterns have shifted. We believed an agent could compress this analysis from days to minutes and identify intervention points that actually move the needle.
What we tested
We built an agent system with four specialised components to diagnose FTFR bottlenecks in real-time. The system processes completed service orders, technician data and parts inventory to identify why jobs require return visits.
Job Pattern Analyser: Ingests completed service orders from the field service platform. Extracts job type, asset age, customer tier, initial symptom description and completion status. Tags each job with complexity markers based on historical patterns. Identifies which job characteristics correlate with incomplete resolutions.
Technician Performance Profiler: Processes technician assignment data, certification records and historical completion rates. Builds capability profiles showing each technician's strengths across job types, asset categories and complexity levels. Identifies skill-job mismatches that lead to incomplete fixes.
Parts Availability Correlator: Analyses parts usage patterns against inventory levels and supplier lead times. Identifies which jobs fail due to missing parts versus diagnostic errors. Tracks seasonal demand patterns and supplier reliability by part category.
FTFR Impact Quantifier: Combines insights from the three analysis components to rank intervention opportunities by potential FTFR improvement. Calculates the revenue impact of each bottleneck and recommends specific actions: training programs, inventory adjustments or dispatch rule changes.
We tested this system against 2,847 completed service orders from a mid-market HVAC operation over 6 weeks. The dataset included 23 technicians, 1,200 unique assets and 340 distinct part numbers.
The architecture
FSM SYSTEM AGENT PIPELINE
┌─────────────────┐ ┌──────────────────────────┐
│ Completed │ │ [1] Job Pattern │
│ Service Orders │────────→ │ Analyser │
│ │ │ │
│ Technician │ └─────────┬────────────────┘
│ Records │ │
│ │ ┌─────────▼────────────────┐
│ Parts Inventory │────────→ │ [2] Technician │
│ │ │ Performance Profiler │
│ Asset History │ └─────────┬────────────────┘
└─────────────────┘ │
┌─────────▼────────────────┐
│ [3] Parts Availability │
│ Correlator │
└─────────┬────────────────┘
│
┌─────────▼────────────────┐
│ [4] FTFR Impact │
│ Quantifier │
└─────────┬────────────────┘
│
ACTIONABLE
INSIGHTS
│
┌──────────────────┼──────────────────┐
│ │ │
▼ ▼ ▼
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ Training │ │ Inventory │ │ Dispatch │
│ Programs │ │ Adjustments │ │ Rules │
└─────────────┘ └─────────────┘ └─────────────┘
│ │ │
└──────────────────┼──────────────────┘
│
▼
┌─────────────────┐
│ FSM SYSTEM │
│ Updates │
└─────────────────┘
What worked
The agent successfully identified three specific bottlenecks causing the FTFR plateau, with quantified impact for each intervention area.
- Identified that 31% of repeat visits stemmed from assigning junior technicians to complex chiller repairs - specific enough to redesign dispatch rules
- Quantified parts stockout impact: 18% of incomplete jobs involved 6 specific part numbers with 3+ week lead times
- Discovered seasonal patterns: FTFR dropped 8 points every summer when demand spiked but dispatch logic remained static
- Correlated technician certification gaps with job failure rates - found 4 technicians needed refrigerant handling training
- Calculated ROI for each intervention: parts inventory expansion would cost £12K but prevent £47K in repeat visit costs
- Could not analyse customer communication quality as a FTFR factor - field service platform doesn't capture interaction notes consistently
- Struggled with asset condition assessment - technician condition ratings were subjective and inconsistent
- Missed travel time impact on job quality - tired technicians after long drives showed lower FTFR but system couldn't correlate fatigue
What we learned
- FTFR stagnation has measurable causes. Operations directors often treat first-time-fix rates as mysterious. But skill mismatches, parts availability and complexity routing errors can be quantified. The agent found that 67% of repeat visits traced to just three identifiable factors.
- Intervention prioritisation requires cost-benefit analysis. Not all FTFR improvements are worth pursuing. Training 4 technicians in refrigerant handling would improve FTFR by 6 points and cost £3K. Expanding parts inventory would improve FTFR by 3 points but cost £12K. The agent ranked interventions by impact per pound invested.
- Seasonal patterns matter more than annual averages. FTFR averaged 73% annually but ranged from 81% in winter to 65% in peak summer. The agent identified that summer dispatch logic needed different rules - assigning only senior technicians to emergency calls when demand peaked.
- Field service platform integration multiplies insight value. The agent analysis is useful, but the real impact comes from updating dispatch rules, technician assignments and parts ordering directly in the field service system. Without integration, insights become reports that gather dust.
Potential for our clients
Experiment status
The FTFR analysis system is production-ready for field service operations with structured job data and technician records. The agent requires 6 weeks of historical service orders to establish baseline patterns. Integration with field service platforms works through standard APIs for job data, technician records and parts inventory.
Operations directors struggling with stagnant FTFR should consider this approach. Typical engagement scope involves 2 weeks to configure the analysis system and 4 weeks to implement the highest-impact interventions. This maps to a Workflow Sprint where we analyse existing FTFR bottlenecks and implement specific dispatch rule improvements.