How do I set up a maintenance dispatch platform for after-hours emergency repairs and photo-based ticket triage?

Here’s a practical way to set up a maintenance dispatch platform for after-hours emergency repairs and photo-based ticket triage.

1) Define the workflow first

Map the lifecycle of a request:

1. Request received
   Phone call, text, web form, QR code, tenant portal, or chatbot.
2. Auto-triage
   Categorize by urgency, location, asset, and availability of photos.
3. Emergency decision
   Decide if it’s:
   • true emergency
   • urgent but can wait until morning
   • non-urgent
4. Dispatch / assign
   Route to on-call technician, contractor, or supervisor.
5. Track completion
   Status updates, photos, notes, timestamps, invoicing.
6. Closeout and audit
   Ensure SLA compliance and after-hours cost reporting.

---

2) Choose the core system architecture

You typically need four components:

A. Intake layer

Where requests enter the system:

• Web form
• Mobile app
• SMS/text
• Email ingestion
• Call center/IVR
• QR code on-site

B. Triage engine

Rules-based or AI-assisted engine that:

• scores urgency
• identifies likely issue type from text/photos
• checks building/asset context
• decides next action

C. Dispatch engine

Routes cases based on:

• on-call schedule
• technician skill
• geography
• workload
• SLA priority
• contractor availability

D. Field response / technician app

For the after-hours responder:

• assignment alerts
• maps and access notes
• issue history
• ability to upload photos
• job status and time logging

---

3) Build the after-hours emergency logic

Create explicit rules so the system doesn’t guess badly.

Example emergency categories

• Active water leak / flooding
• Fire, gas, electrical hazard
• Elevator entrapment
• No heat in freezing conditions
• Break-in / security breach
• Sewer backup
• Critical equipment failure affecting operations

Example triage rules

• If keywords indicate immediate hazard, mark Emergency
• If photo shows water pooling, exposed wiring, smoke, or broken entry, escalate
• If issue is in a critical asset area, increase priority
• If no photos but message suggests life/safety risk, still escalate
• If it’s cosmetic or routine, defer until business hours

Decide response levels

• Level 1: Immediate dispatch
• Level 2: Call back within 15 minutes for clarification
• Level 3: Schedule next business day
• Level 4: Auto-close or route to normal queue

---

4) Add photo-based triage

Photos can dramatically improve triage, but only if structured well.

What to ask for

Require requesters to upload:

• 1–3 photos minimum
• wide shot + close-up
• location label
• optional short video for active leaks or noise issues

How to triage photos

You can use:

• Human review for highest accuracy
• AI image classification to detect water, fire, damage, broken fixtures, etc.
• OCR to read equipment labels or fault codes
• Vision + text combined scoring for better prioritization

Best practices

• Ask for photos in a guided flow
• Auto-compress and store safely
• Timestamp and geotag if allowed
• Keep a review queue for uncertain cases
• Never rely on photo-only decisioning for life/safety risks

---

5) Set up on-call scheduling and escalation

This is the backbone of after-hours dispatch.

What you need

• Rotating on-call calendar
• Backup/on-call secondary
• Escalation ladder if no response
• Coverage by region and skill type

Example escalation

1. Notify primary on-call technician
2. If no response in 5 minutes, alert backup
3. If no response in 10 minutes, notify supervisor
4. If still unresolved, call contractor/vendor or emergency services if needed

Integrations

• Calendar system
• SMS/push notification
• Phone call escalation
• Slack/Teams alerts for supervisors

---

6) Build a rules engine for routing

Use a decision matrix combining:

• Urgency
• Property type
• Asset type
• Technician skill
• Distance/travel time
• Current workload
• Contract/SLA obligations

Example:

A leaking boiler in a senior housing facility at 2 a.m.:

• urgency = critical
• asset = heating
• location = senior housing
• route = immediate dispatch to HVAC on-call with supervisor alert

---

7) Design the data model

At minimum, track:

• Ticket ID
• Requester info
• Property/site
• Unit/room/location
• Issue category
• Priority
• Status
• Photos/videos
• Timestamp
• Assigned technician
• On-call escalation events
• Actions taken
• Resolution notes
• Cost/time/labor
• SLA metrics

---

8) Choose your technology stack

If you’re building from scratch:

Simple stack

• Frontend: React / Next.js
• Backend: Node.js / Python / .NET
• Database: PostgreSQL
• File storage: S3-compatible object storage
• Notifications: Twilio + email + push
• Workflow engine: Temporal / Camunda / custom rules service
• AI services: vision model API + text classification API

Faster no-code / low-code option

• Intake forms: Typeform, Jotform, Microsoft Forms
• Workflow automation: Zapier, Make, n8n, Power Automate
• Ticketing: ServiceNow, Zendesk, Freshservice, Jira Service Management
• Dispatch: custom add-on + SMS/Teams integrations

If you need this working quickly, a ticketing platform plus automation tools is usually the fastest route.

---

9) Make the request form triage-friendly

Keep the form short, but smart.

Required fields

• Name and contact
• Site/unit
• Issue type
• “Is anyone in danger?”
• “Is there active water, smoke, gas smell, power outage, or security issue?”
• Photo upload
• Best callback method

Smart branching

If user selects:

• Water leak → ask if water is actively flowing and where it’s coming from
• Electrical issue → ask if there is smoke, sparking, or exposed wire
• Heat outage → ask for temperature and affected area
• Security issue → ask if police or emergency services are needed

---

10) Define business hours vs after-hours behavior

The platform should behave differently depending on time.

Business hours

• normal queue
• supervisor review
• standard SLA routing

After-hours

• emergency screening
• immediate dispatch only for critical issues
• use scripted callback questions
• minimize false alarms
• log all decisions carefully

---

11) Create technician playbooks

For each emergency type, write a playbook:

• What qualifies
• What to ask on the phone
• First response steps
• Safety precautions
• Who to notify
• When to escalate externally

This keeps response consistent and reduces liability.

---

12) Add reporting and KPI dashboards

Track:

• time to acknowledge
• time to dispatch
• time to arrival
• resolution time
• emergency rate
• false positive rate
• after-hours cost
• repeat incident rate
• photo triage accuracy
• SLA breaches

These metrics tell you whether your triage rules are too strict or too loose.

---

13) Security and compliance

Because you’re handling property, tenant, and possibly personal data:

• Role-based access control
• Audit logs
• Secure media storage
• Data retention policy
• Encryption in transit and at rest
• Consent/notice for photo collection
• Emergency contact verification

If you serve regulated environments, review relevant housing, privacy, and safety requirements.

---

14) Recommended implementation path

Phase 1: MVP

• ticket intake form
• photo upload
• rules-based emergency routing
• SMS/email on-call alerts
• manual dispatcher dashboard

Phase 2: Automation

• on-call schedule integration
• escalation ladder
• SLA timers
• technician mobile workflow
• reporting dashboard

Phase 3: AI assistance

• photo classification
• text severity scoring
• duplicate detection
• auto-summarization for dispatchers

---

15) Practical starter workflow

A good first version might look like this:

1. Tenant submits issue with photos
2. System asks a few emergency questions
3. Rules engine scores ticket
4. If critical, SMS/call on-call technician
5. Dispatcher sees summary + images
6. Technician accepts job in mobile app
7. Technician updates status and uploads completion photos
8. System closes ticket and logs SLA/cost data

---

If you want, I can also give you one of these:

1. a sample workflow diagram,
2. a feature list for an MVP, or
3. a recommended software stack for building it fast.