Case Studies / Field & Asset Operations
⚡ Field & Asset Operations · Illustrative Scenario

Work order management: triage,
routing, and closeout automated

A facilities management company handling 600+ work orders per month had dispatcher bottlenecks adding more than a business day to 28% of all orders. The Aizen Event revealed that 5 of 8 steps were fully automatable with deterministic rules: and that the highest-value step required only RAG, not expensive probabilistic AI.

71%
Dispatcher time reduction
$410K
Annual savings
12wk
To deployment
600+
Orders/month handled
Step 01 · Current State

The workflow before the Aizen Event

8 steps spanning intake to closeout, 600+ work orders per month across 12 sites. Dispatcher bottlenecks and manual processes delayed ~28% of orders by >1 day. High billing error rate and parts-ordering inefficiency.

#
Workflow step
Avg. time
Performed by
01
Work order intake & classification
Customer submits request online or by phone. Dispatcher manually categorizes, prioritizes, and assigns initial fields.
25 min
per order
Dispatcher
02
Asset history lookup
Dispatcher + technician manually search asset maintenance history, prior work orders, and known failure patterns
30 min
per order
Dispatcher + Technician
03
Technician assignment
Dispatch supervisor manually matches work order to available technician based on skills, location, current workload
20 min
per order
Dispatch supervisor
04
Parts inventory check
Parts coordinator manually searches inventory across 3 warehouses, flags shortages, identifies substitutes
15 min
per order
Parts coordinator
05
On-site diagnosis
Field technician conducts assessment. Cannot be automated. Output drives repair strategy.
variable
on-site time
Field technician (Human Required)
06
Work documentation
Technician manually completes forms detailing fault codes, parts used, time spent, and follow-up actions
45 min
per order
Field technician
07
Parts ordering & procurement
Parts coordinator manually creates POs, submits to vendor, tracks deliveries. Manual errors common.
20 min
per order
Parts coordinator
08
Billing & closeout
Billing staff manually generates invoice, applies billing codes, closes work order in CMMS. Error rate: 8%.
30 min
per order
Billing staff
TOTAL LABOR PER ORDER
~3.1 hours
MONTHLY ORDERS
600+
DELAYED >1 DAY
28%
Step 02 · Aizen Event

Classifying every step.

In the Aizen Event, we map each workflow step to an AI type: Deterministic (rules with predictable output), RAG (retrieval-grounded), Probabilistic (learned judgment), or Human Required. Classification drives build vs. defer decisions.

Step AI type Recommendation New state Time saved
01
Work order intake & classification
Manual categorization and routing
 Deterministic Categorize by work type (preventive/corrective/emergency), priority level, and asset class from work order fields. Route to appropriate queue automatically. Emergency orders page on-call technician. ● Automated 22 min
02
Asset history lookup
Manual search of prior records
 RAG Retrieve asset maintenance history, prior work orders, warranty status, and known failure patterns. Present to technician before dispatch as pre-visit brief on mobile app. ● Augmented 24 min
03
Technician assignment
Manual supervisor matching
 Probabilistic Match work order (type, location, required certification) to available technicians. Skills matrix + proximity routing + current workload scoring. Supervisor approves or selects alternatively. ● Augmented 14 min
04
Parts inventory check
Manual warehouse searches
 Deterministic API query to inventory system on work order creation. Check stock across all 3 warehouses, flag shortages, identify substitutes. Pre-populate parts list for technician. ● Automated 13 min
05
On-site diagnosis
Field technician assessment
 Human Required Field technician assessment. Cannot automate. AI provides structured pre-visit brief (asset history, known failure modes, recommended tools) via mobile app to accelerate on-site decision-making. ○ Unchanged —
06
Work documentation
Manual form completion by technician
 Probabilistic Technician answers structured prompts via mobile app (or dictates). AI generates formatted work order completion notes, tags fault codes, recommends follow-up actions. Tech reviews and submits. ● Augmented 32 min
07
Parts ordering
Manual PO generation and submission
 Deterministic Auto-generate PO from work order parts list. Route for approval if >$500. Send to preferred vendor automatically. Update inventory on confirmation. Zero manual PO work. ● Automated 18 min
08
Billing & closeout
Manual invoice generation and codes
 Deterministic Auto-generate invoice from work record (labor hours + parts). Apply correct billing codes by work type and client contract. Close work order in system. Error rate drops from 8% to <1%. ● Automated 27 min
DeterministicRules-based, predictable output
RAGRetrieval-augmented, document-grounded
ProbabilisticContextual judgment, human oversight required
Human RequiredDo not automate: field expert judgment
Step 03 · Priority Matrix

What to build first. What to defer.

We plot each step by implementation complexity (X) against business impact (Y). Steps in the top-left are Quick Wins: start here. Top-right are Strategic Investments: build after wins are live. Bottom-right: skip entirely.

QUICK WIN STRATEGIC INVESTMENT AUTOMATE LATER AVOID / HUMAN ONLY IMPLEMENTATION COMPLEXITY → BUSINESS IMPACT → 1 2 3 4 5 6 7 8 BUILD ORDER: 1,4,7,8 → 3 → 2,6
QUICK WIN (start here)
1
Intake classification

Rules-based. First build.

4
Parts inventory

API call. Phase 1.

7
Parts ordering

Template automation.

8
Billing/closeout

Fully rule-driven.

STRATEGIC INVESTMENT
6
Work documentation

Probabilistic with mobile prompts. Phase 2.

2
Asset history RAG

Corpus indexing. Phase 2.

5
On-site diagnosis

Field expert judgment. Not touched.

Step 04 · New State

What happened to each step.

The redesigned workflow after the Aizen Event implementation. Every step with a new state shows the before design, what changed, and who now handles what.

01
Before
Intake & classification
Dispatcher manually categorizes, prioritizes, and routes. 25 min per order.
● Automated
Auto-classification engine
Work order fields auto-classify: work type, priority (P1–P4), asset class, building/site. Routed to correct queue instantly. Emergency orders page on-call technician automatically.
02
Before
Asset history lookup
Dispatcher + technician manually search prior records. 30 min per order.
● Augmented
RAG asset brief
RAG retrieves full asset profile: maintenance history, prior fault codes, warranty status, manufacturer notes, failure pattern analysis. Packaged as pre-visit brief delivered to technician's mobile app before arrival on site.
03
Before
Technician assignment
Supervisor manually matches technician to order. 20 min per order.
● Augmented
Assignment model
Model scores all available technicians against work order requirements: certification match, current proximity, active workload, client relationship history. Top recommendation presented; supervisor confirms or selects alternative. Time: ~6 min vs 20 min.
04
Before
Parts inventory check
Coordinator manually searches 3 warehouses. 15 min per order.
● Automated
Inventory API query
On work order creation, system queries inventory across 3 warehouses. Pre-populates parts list with stock status and location. Flags lead times for out-of-stock items. Zero coordinator involvement for standard orders.
05
Before
On-site diagnosis
Technician conducts site assessment. Variable time based on job complexity.
○ Unchanged
Diagnosis (optimized)
Field technician conducts assessment. AI-powered mobile app provides asset brief and structured fault reporting prompts, but diagnosis decision is entirely the technician's. Pre-visit intel reduces on-site time by avg 12 min.
06
Before
Work documentation
Technician manually fills out forms. 45 min per order.
● Augmented
Mobile prompts + AI summary
Technician answers 8 structured prompts via mobile app (or dictates). AI generates complete work order completion notes, applies ASHRAE fault codes, flags any warranty-relevant findings, recommends preventive follow-up. Tech reviews and submits. Time: ~13 min vs 45 min.
07
Before
Parts ordering
Coordinator manually creates POs and submits. 20 min per order. 8% error rate.
● Automated
Auto-generated POs
PO auto-generated from parts list and preferred vendor catalog. Orders <$500 auto-approved and transmitted. Orders >$500 routed for supervisor approval. Inventory updated on confirmation. Error rate drops to <1%.
08
Before
Billing & closeout
Billing staff manually generates invoices and codes. 30 min per order. 8% error rate.
● Automated
Auto-invoice generation
Invoice auto-generated from work record. Labor hours from field app timestamps, parts from PO records, billing codes from work type taxonomy. Applied to correct client contract. Work order closed in CMMS. Zero manual work.
New avg. labor per order
~0.9 hours ↓ 71%
Orders delayed >1 day
6% ↓ 78%
Step 05 · Investment & ROI

The numbers.

Build cost, deployment timeline, and three-year return: based on 600+ monthly orders across 12 sites, blended field service labor rates, and measured time savings post-deployment.

Investment breakdown
Discovery sprint + asset corpus indexing plan$28,000
Phase 1 build (intake + parts + ordering + closeout automations)$68,000
Phase 2 build (RAG asset history + documentation AI + assignment model)$112,000
Mobile app integration + field testing$32,000
Total investment$240,000
Year 1 labor savings
$410K
Dispatcher + coordinator time savings across 600 orders/mo + reduced billing errors + faster parts ordering. Based on 7,200 orders/yr × 2.2 hrs saved × $25.50/hr blended rate.
Payback period
7.0 mo
Full investment recovered within 7 months of live deployment (deployment at week 12 of engagement).
3-year cumulative ROI
$990K
Net of full investment. Includes conservative 8% annual order growth and system maintenance at $22K/yr. Does not include first-time fix rate improvements or response time gains.
Secondary outcomes
Emergency work order response time dropped from avg 4.1 hrs to 1.8 hrs
First-time fix rate improved from 71% to 84% (better pre-visit asset intelligence)
Parts ordering errors down 91% (auto-generated POs vs manual)
28% of orders delayed >1 day reduced to 6%
// Illustrative scenario note

This case study represents a composite of typical field service and facilities management engagements. Metrics are based on observed industry benchmarks and comparable implementations. All investment figures are illustrative ranges based on the Otonmi service tiers.

Other case studies
Federal Government FOIA: 45-day cycle → 17-day average Federal agency, 3,200+ annual requests → Enterprise Portfolio reporting: 8 hrs → 40 min Wealth management, 42 RMs → Healthcare Patient intake: 3 FTEs eliminated Healthcare services, 400+ intakes/week →

Run this on your workflow.

Tell us about the process your team is spending the most time on. We'll classify every step and tell you what's worth building: and what isn't.

Book a workflow session → All case studies