Home / Sectors / On-Demand
Vertical · On-Demand Platforms

Matching, dispatch,
delivered to the second.

Ride-hailing, food delivery, home services, on-site repair, every on-demand category lives or dies on real-time matching, ETA honesty and surge logic that customers tolerate. We engineer the dispatch engine, driver and provider apps, and the payment flows behind them to perform under Riyadh-traffic, payday-spike reality.

<200msMatch decision latency p95
±2 minETA accuracy across categories
Nafath + QiwaProvider compliance native
Service request
Service & locationGeolocated
TimeNow or scheduled
Step 1Request

Customers request a service by location and time in a few taps.

Step 2Match & dispatch

The nearest available provider is matched and dispatched live.

Service completed

Rated, paid and settled, all in-app.

Step 3Complete & pay

Service is completed, rated and paid, settled in-app.

Swipe to explore
On-Demand, engineered

From request to rated service, matched in seconds.

Instant requests, provider matching, live ETA, in-app pay and ratings, one platform built end to end.

9:415G
ServicesNear you
Home cleaningfrom
80
AC repairfrom
150
Plumbingfrom
120
Request
ServiceCleaning
LocationHome
TimeNow
Matching
Provider
Matched
ETAMinutes12
Rating4.8
Payment
Service120
Fee10
Total130 SAR
Completed
Service Completed
Rated & paid in-app
RequestsLive matchingETA trackingIn-app payRatings
The challenge

Where do on-demand platforms fail?

The architectural shortcuts that look fine in a pilot and become structural problems at city scale.

Matching latency

Match decisions take seconds during peak; riders cancel, drivers idle, both sides churn. Latency is the unit economic, not a UX detail.

ETA dishonesty

Optimistic ETAs drive the cancel rate. Honest ETAs powered by traffic, driver state and historical patterns drive retention.

Provider compliance drift

Driver/provider licences and Qiwa status expire silently. The platform discovers it when a regulator does.

Surge mispricing

Static surge multipliers anger users without serving real supply scarcity. Demand modelling is engineered, not heuristic.

Payment splits

Per-trip payouts to drivers, restaurants, partners, tax, split incorrectly once and the support cost compounds for months.

Trust & safety blind spots

No SOS, no journey monitoring, no reviewable evidence after an incident. One viral story is enough to stall growth.

What we build

What do we engineer?

Production modules across ride-hailing, delivery and home-services engagements.

Matching & dispatch engine

Sub-200ms matcher with geo-sharded supply pools, retry handling and explainable assignment for support and audit.

Honest ETA service

Live traffic, road-network state and historical demand feed an ETA model calibrated to the corridor, not the average.

Surge & pricing model

Demand-supply elasticity with transparent surge signalling, cooldowns and price-ceilings that protect customer trust.

Driver / provider app

Battery-aware, low-data, offline-tolerant native app with Nafath onboarding, document expiry and earnings transparency.

Payment splits & payouts

Per-trip ledger postings, multi-party splits, ZATCA-compliant invoicing and reconciled payout flows for drivers and partners.

Trust & safety stack

SOS, journey sharing, anomaly detection on routes and timings, plus a case-management surface for the safety team.

<200msMatch decision latency at p95
±2minETA accuracy in the served corridors
−34%Cancel rate post-ETA & surge re-engineering
100%ZATCA-compliant per-trip e-invoicing
Common questions

Engineering Saudi on-demand, answered

Do you build for ride-hailing, delivery or services?

All three. The matching, dispatch, payments and safety modules are the same architecture, what changes is the supply-side onboarding, asset types and the trip lifecycle. Cross-category platforms benefit most from this.

Can the platform handle a city-wide event?

Yes, we engineer for surge, not for steady state. Geo-sharded matchers, async pricing recompute and elastic infra mean the platform degrades gracefully rather than collapsing on the night that matters.

How do you handle TGA, Qiwa and Nafath?

Provider onboarding is wired into Nafath identity and Qiwa employment status, with TGA category checks where required. Document expiry monitoring keeps the active fleet compliant, automatically.

What about per-trip e-invoicing?

Each completed trip generates a ZATCA Phase 2-compliant e-invoice automatically, with credit notes for cancellations and refunds. No manual issuance, no end-of-month surprises.

Engineer on-demand
that holds at city scale.

Tell us where matching, pricing or payouts strain today. We'll map the engine that handles peak day without paging the founders.

We usually reply within one business day
Other sectors

Adjacent verticals we engineer