Welcome

We blend strategy, design and engineering to build intelligent systems that help businesses in Saudi Arabia and the GCC grow and stand out.

Jiwar

Municipal reports and field operations, from a citizen photo to a verified fix

  • A Shfrah product
  • Government & municipalities
  • AI
  • Mobile
  • At a glance

Sector

Government & municipalities

Platforms

Web · iOS · Android · WhatsApp

Modules

9

Integrations

8

Screens

4

Total duration

24 weeks

Jiwar takes citizen reports from the app, WhatsApp and web with a photo and a location resolved to the National Address. AI classifies the photo and the text, merges nearby duplicates and routes each report to the right department or contractor, with an SLA set for its category. Field crews work from their own app with before-and-after evidence, citizens rate the fix, and leadership follows a live district heat map, SLA compliance and contractor performance.

Jiwar: screen 1
Operations centre: a heat map of open reports by district, SLA compliance and the districts falling behind.
  • Challenge

  1. 01

    One issue is reported dozens of times across channels, causing repeat site visits and scattered crew effort.

  2. 02

    Triage and routing are manual: staff read each description and pick an owner, so reports bounce between departments and contractors for days.

  3. 03

    There is no verified proof of work: reports close without an after photo, and citizens report the same issue again.

  4. 04

    Leadership has no live view of districts or of each contractor against its contract, so decisions rest on late monthly reports.

Who it is built for

  • Citizen or resident

    Reports in under a minute with a photo and location, and knows when it will be fixed

  • Operations centre agent

    Reports already classified, clustered and routed, ready to confirm

  • Department or contractor supervisor

    A prioritised task list with a deadline for every report

  • Field crew

    A daily route, the exact spot and before-and-after evidence from the phone

  • Executive leadership

    A live district map, SLA compliance and contractor performance

  • Modules

What the system does

Connected modules working on one set of data.

  • 01

    Omnichannel intake

    App, WhatsApp, web and call centre in one inbox, with photo, location and National Address

  • 02

    AI triage

    Classifies photo and text, rates severity and proposes the owner

  • 03

    Duplicate clustering

    Groups similar nearby reports under one master and updates every reporter

  • 04

    Routing & SLAs

    Routing by category, district and contract zone, with an SLA per category and severity and automatic escalation

  • 05

    Field crew app

    Daily tasks and navigation, geotagged before-and-after photos, works offline

  • 06

    Contractor management

    Contract zones, SLAs and penalties, with work verified before each payment certificate

  • 07

    GIS & heat maps

    Heat maps by district and category, with asset layers such as light poles

  • 08

    Citizen feedback

    A rating after closure, and reopening when the issue is not resolved

  • 09

    Executive dashboard

    SLA compliance, backlog, department and contractor performance, with periodic reports

Where the AI works

  • AI · 01

    A vision model classifies the report photo (pothole, broken street light, waste, water leak, damaged sidewalk), rates severity and rejects unusable photos.

  • AI · 02

    An Arabic language model reads descriptions written in local dialects and extracts the category and risk cues such as "near a school" or "constant leak".

  • AI · 03

    Duplicate detection combines distance, category, time and image similarity, attaching a new report to an open one instead of opening a new task.

  • AI · 04

    Before-and-after photo comparison verifies the work before closure, and a forecast of report volume by district and season helps plan crews.

  • Screens

The system as the client's team and customers use it.

Jiwar: screen 2
Report detail: AI classification and severity, the duplicate cluster within 60 metres, and the proposed assignment.
Jiwar: screen 3
Citizen app: a photo, a location on the National Address and automatic classification, with a notice when a similar report already exists.
Jiwar: screen 4
Field crew app: the task, its deadline and location, with before-and-after evidence the system checks before closure.
  • Architecture

Independent services, in-Kingdom data, and integrations over documented APIs.

Channels

  • Citizen app: iOS & Android
  • WhatsApp bot: Reports with photo & location
  • Web & call centre: Web form & call centre
  • Field crew app: Offline, before & after photos
  • Ops console: Operations & leadership

Edge

  • API gateway: Rate limits & WAF
  • Identity: Nafath, roles & contract zones

Services

  • Intake service: Report intake & normalisation
  • Routing & SLA: Routing, SLAs & escalation
  • Work orders: Tasks, contractors & completion
  • Geo service: Addresses, districts & maps
  • Notifications: WhatsApp, SMS & push

AI

  • Photo classifier: Category, severity & before/after check
  • Arabic text model: Descriptions in local dialects
  • Duplicate detector: Distance, time & image similarity

Data

  • PostgreSQL · PostGIS: Reports, points & districts
  • Kafka: Report events
  • Object storage: Photos, in-Kingdom

Integrations

  • Nafath: National SSO
  • SPL National Address: National & short address
  • Balady: Report sync where applicable
  • Esri ArcGIS: District & asset layers
  • Contractor systems: Work orders & completion

Connections

  • Citizen app to API gateway
  • WhatsApp bot to API gateway
  • Web & call centre to API gateway
  • Field crew app to API gateway
  • Ops console to API gateway
  • API gateway to Identity
  • API gateway to Intake service: REST
  • API gateway to Work orders
  • Intake service to Photo classifier
  • Intake service to Arabic text model
  • Intake service to Duplicate detector
  • Intake service to Geo service
  • Intake service to Routing & SLA
  • Routing & SLA to Work orders
  • Work orders to Notifications
  • Work orders to Photo classifier: after photo
  • Intake service to PostgreSQL · PostGIS
  • Intake service to Kafka: events
  • Intake service to Object storage
  • Identity to Nafath
  • Geo service to SPL National Address
  • Geo service to Esri ArcGIS
  • Intake service to Balady: sync
  • Work orders to Contractor systems: API

Stack

  • Next.js
  • NestJS
  • Flutter
  • PostgreSQL · PostGIS
  • Kafka
  • MapLibre
  • Python · PyTorch
  • Kubernetes
  • Workflow

Report lifecycle

  1. 01

    Report with photo and location

    CitizenApp · WhatsApp · Web

  2. 02

    Identity and address

    AutomaticNafath · SPL National Address

  3. 03

    Classify and rate severity

    AutomaticPhoto classifier · Arabic text model

  4. 04

    Detect and merge duplicates

    AutomaticDuplicate detector

  5. 05

    Route and set the SLA

    AI, then operations centreRouting & SLA

  6. 06

    Assign a crew or contractor

    SupervisorWork orders · Contractor systems

  7. 07

    Fix with before-and-after evidence

    Field crewField crew app

  8. 08

    Verify and close

    AI, then supervisorPhoto classifier

  9. 09

    Notify and collect a rating

    CitizenWhatsApp · App · Balady

  • Data & API

The core entities and relations, and a live example from the API.

Report

  • iduuidPK
  • codetextunique
  • channelenum(app,whatsapp,web,call)
  • category_iduuidFK
  • district_iduuidFK
  • locationgeography(Point)
  • national_addresstext
  • cluster_iduuidFK
  • statusenum

Cluster

  • iduuidPK
  • master_report_iduuidFK
  • category_iduuidFK
  • severitysmallint
  • centroidgeography(Point)
  • reports_countint
  • sla_due_attimestamptz

Category

  • iduuidPK
  • name_artext
  • department_iduuidFK
  • sla_hourssmallint
  • severity_rulesjsonb

District

  • iduuidPK
  • name_artext
  • city_codetext
  • boundarygeography(Polygon)
  • contract_zone_iduuidFK

WorkOrder

  • iduuidPK
  • cluster_iduuidFK
  • assignee_typeenum(crew,contractor)
  • assignee_iduuid
  • due_attimestamptz
  • statusenum
  • closed_attimestamptz

Evidence

  • iduuidPK
  • work_order_iduuidFK
  • kindenum(before,after)
  • storage_keytext
  • taken_attimestamptz
  • locationgeography(Point)
  • match_scorereal

Contractor

  • iduuidPK
  • nametext
  • cr_numbertext
  • contract_notext
  • zone_idsuuid[]
  • penalty_rulesjsonb

Feedback

  • iduuidPK
  • report_iduuidFK
  • ratingsmallint
  • commenttext
  • reopenedboolean
  • attimestamptz
  • Category 1—∞ Report
  • District 1—∞ Report
  • Cluster 1—∞ Report
  • Cluster 1—1 WorkOrder
  • Contractor 1—∞ WorkOrder
  • WorkOrder 1—∞ Evidence
  • Report 1—1 Feedback
POST /v1/reports Open a report from any channel

Request

{
  "channel": "app",
  "reporter": { "nafath_sub": "nfd_7Q2K9", "mobile": "+9665•••••381" },
  "location": { "lat": 24.7368, "lng": 46.8225, "accuracy_m": 6 },
  "description": "Large pothole in front of the school gate; cars are swerving around it",
  "photos": ["upl_91c2", "upl_91c3"]
}

Response

{
  "id": "rpt_01JD4R8W2M",
  "code": "JWR-26-118180",
  "status": "routed",
  "classification": {
    "category": "pothole",
    "confidence": 0.97,
    "severity": "high",
    "risk_cues": ["near_school"]
  },
  "national_address": { "short": "RNSB4729", "district": "Al Naseem" },
  "duplicate_of": null,
  "assigned_to": { "type": "contractor", "zone": "EAST-2" },
  "sla_due_at": "2026-11-08T18:42:00+03:00",
  "tracking_url": "/t/118180"
}

Webhook event · report.closed

{
  "event": "report.closed",
  "cluster": "CLS-26-020931",
  "master": "JWR-26-118180",
  "reporters_notified": 6,
  "closed_by": "crew_asphalt_7",
  "after_photo_match": 0.93,
  "within_sla": true,
  "at": "2026-11-08T16:24:10+03:00"
}
  • Integrations

  • Nafath

    Citizen sign-in with the national SSO

  • SPL National Address

    Coordinates to National Address and short address

  • Balady

    Report and status sync with the Balady platform where applicable

  • WhatsApp Business

    Intake with photo and location, plus status updates

  • Esri ArcGIS

    District and asset layers and heat maps

  • Contractor systems

    Work orders out and completion back over API

  • Call centre (CTI)

    Phone reports opened with caller details

  • SMS gateway

    Text updates for people without the app

  • Security & compliance

  • Data and photos hosted inside the Kingdom, in line with government data classification.
  • National Cybersecurity Authority Essential (ECC) and Cloud (CCC) Cybersecurity Controls.
  • Personal Data Protection Law (PDPL) alignment: faces and number plates blurred automatically before photos are shown.
  • Reporter identity and number are never shown to contractors or field crews.
  • Role- and zone-based access: each contractor sees only the reports in its contract zone.
  • Encryption in transit and at rest with managed keys.
  • Immutable audit trail for every assignment, status change and closure.
  • Delivery plan

Clear phases with measurable deliverables.

  1. 01

    Discovery & process mapping

    • Workshops with the operations centre, technical departments and contractors
    • Inventory of categories, SLAs and contract zones
    3 weeks
  2. 02

    UX & architecture

    • Citizen and crew app prototypes tested in the field
    • Spatial data, routing and API design
    3 weeks
  3. 03

    MVP build

    • Omnichannel intake, routing and SLAs
    • Citizen app, crew app and operations console
    8 weeks
  4. 04

    Integrations & GIS

    • Nafath, National Address, Balady and WhatsApp
    • District and asset layers and contractor systems
    4 weeks
  5. 05

    Model training & pilot

    • Train photo and text models on labelled historical reports
    • Pilot in a small set of districts
    4 weeks
  6. 06

    Go-live & hypercare

    • Phased launch by district and category
    • Training for field crews and the operations centre
    2 weeks
Total duration
24 weeks
Team
Project manager, business analyst, UX designer, a GIS engineer, 3 backend and frontend engineers, 2 mobile engineers, an ML engineer, QA and DevOps.

Performance

  • < 2 min Automatic triage and routing
  • 85%+ Duplicates caught before assignment
  • 100% Reports closed with an after photo
  • 99.9% Availability

Related work

All work

Have a similar project?

Start a project
Contact us