Before the Route: Why AI Route Optimisation in the Middle East Starts with the Address

AI route optimisation in the Middle East is not simply a routing problem. In the GCC, the first operational challenge is location confidence: can the system understand where the customer actually is before it calculates the most efficient route?

An e-commerce fulfilment team in Dubai plans 1,200 delivery routes for the day. The optimisation engine is capable. The fleet is available. By noon, a meaningful share of deliveries has failed — not because of traffic, driver capacity or dispatch discipline, but because the addresses point to the wrong buildings, the wrong compound gates, or nowhere a driver can reliably reach.

Every supply chain leader running last-mile operations in the GCC has a version of this problem. The technology stack appears to work. The dispatch plan looks efficient. The economics still miss the business case.

In most global markets, the route is the hard problem. In the GCC, the address is often the hard problem. The route comes later.

AI-powered route optimisation only delivers its full cost benefit in the Middle East when it is built on a location-intelligence layer that understands how Emirati, Saudi and Qatari addresses actually work — Makani codes in the UAE, the Saudi National Address system, Qatar’s zone-based addressing, and bilingual Arabic-English address inputs.

The algorithm matters. The data layer beneath it matters more.

According to the World Bank’s Logistics Performance Index, the UAE ranks among the top ten logistics performers globally — a signal that the region’s physical logistics infrastructure has advanced faster than many of the software layers sitting on top of it. For supply chain leaders in the GCC, the opportunity is to close that software gap deliberately: improve address confidence, automate dispatch decisions, raise on-time delivery performance and reduce cost-to-serve at scale.

Quick answer

AI route optimisation in the Middle East works when routing, dispatch and visibility are built on accurate location intelligence. For GCC operators, that means validating delivery points, parsing Arabic and English addresses, planning around local constraints, automating fleet and carrier allocation, and learning from every completed delivery through AI route optimization software.

Improve GCC route density with AI-powered planning

Plan around address ambiguity, traffic, heat and delivery windows with route optimisation built for real-world execution.

Explore Route Optimization ?

By the Numbers: AI Route Optimisation in GCC and Middle East Logistics

The business case for AI route optimisation in the Middle East is becoming more measurable as logistics teams move from manual planning to automated routing, dispatch and telematics-led execution.

Metric	Reported benchmark	Why it matters
Fuel savings from AI-enabled route optimisation and dispatching	10–15%	Reduces direct fleet operating cost and emissions exposure.
Reduction in kilometres driven per delivery	12–20%	Improves route density, fleet utilisation and driver productivity.
On-time delivery improvement for GCC fleets using AI route optimisation	From 75–82% to 92–97%	Shows the service-level upside of planning around real constraints.
Route planning time for 50 trucks	From 2–3 hours per day to 5–10 minutes	Converts dispatcher time from manual sequencing to exception control.
AI-based route optimisation impact in UAE logistics	Around 15% lower travel time	Demonstrates the value of real-time planning around traffic, cultural events and local operating patterns.
MENA IT spending forecast	USD 169 billion in 2026	Indicates the wider digital infrastructure investment environment supporting AI adoption.

These benchmarks should not be treated as guaranteed outcomes for every fleet. Actual ROI depends on baseline route density, address quality, fleet mix, carrier fragmentation, order volume, dispatcher maturity and how deeply AI is integrated into execution workflows.

What Is AI Route Optimisation in the Middle East?

AI route optimisation uses machine learning, constraint-based algorithms and real-time operational data to determine the best delivery sequence, vehicle assignment, driver allocation, delivery timing and carrier selection for every order.

In the Middle East, the concept has to go beyond generic map-based routing. A useful system must account for:

• Descriptive and bilingual addresses
• Rapidly expanding residential districts
• Dense city traffic in Dubai, Riyadh, Jeddah, Doha and Abu Dhabi
• Extreme heat and driver welfare constraints
• Prayer-time and Ramadan delivery windows
• Fragmented carrier networks across owned fleets, 3PLs and courier partners
• High customer expectations for same-day, next-day and narrow-window delivery
• Customer promise management through time-slot management

Technical route optimisation is often framed as a vehicle routing problem, or VRP. NVIDIA’s route optimisation workflow describes how routing systems evaluate constraints such as vehicle capacity, service windows, travel time and delivery priorities to produce more efficient fleet movement. In GCC last-mile operations, those constraints must be expanded to include regional realities — especially address ambiguity and execution variability.

That is why AI route optimisation in the Middle East should be understood as a decision system, not just routing software. It connects location intelligence, route planning, dispatch execution and field feedback into one operating loop. In mature operations, this also includes dynamic route planning, where delivery sequences can be adjusted continuously as orders, traffic, driver status and customer availability change.

Editorial Methodology: How to Evaluate AI Route Optimisation Claims

This article evaluates AI route optimisation in the Middle East through an operational lens, not as a generic software category. The analysis focuses on five criteria that matter in GCC last-mile execution:

1. Location confidence: whether the system can resolve ambiguous, bilingual and landmark-based addresses before dispatch.
2. Regional constraint handling: whether heat, prayer windows, Ramadan demand compression, city traffic and delivery preferences are embedded into route planning.
3. Execution readiness: whether routing decisions connect to dispatch automation, driver workflows, carrier allocation and exception management.
4. Learning loops: whether geocoding, ETA and failed-delivery prediction improve from completed deliveries and field feedback.
5. Business impact: whether the platform improves measurable KPIs such as first-attempt delivery, kilometres per drop, cost-to-serve, on-time delivery, WISMO volume and SLA adherence.

Where statistics are included, they are linked to their source. Where source data is unavailable or not directly comparable across fleets, this article avoids unsupported ROI claims.

The GCC’s Real Last-Mile Problem Is an Address Problem

Enterprise e-commerce operations across the GCC share a structural reality that does not exist in the same form in Europe or North America: customer-entered addresses are often descriptive rather than structured, and many geocoding layers were not trained on how people in the region actually describe where they live.

The exact pattern varies by country, but the underlying issue is consistent:

Market	Addressing reality	Last-mile impact
UAE	Many customer-entered addresses remain descriptive — “Villa 23, behind Spinneys Jumeirah, near the blue building.” Makani, Dubai’s 10-digit geolocation standard, works well where adopted, but adoption is inconsistent across customer-facing forms.	Poor coordinate confidence, wrong gate selection, avoidable driver calls and failed first attempts.
Saudi Arabia	The Saudi National Address system and Short Code format are well designed, with strong coverage in Riyadh and Jeddah. Rapid urban expansion means new compounds, towers and residential developments can lag in geocoding databases.	Higher geocode mismatch in new districts, more manual dispatcher intervention and weaker route density.
Qatar	Zone-based addressing — zone, street, building — works well in central Doha but thins out in outer zones.	Driver phone calls before delivery remain common, particularly where building-level precision is weak.
Arabic–English parity	Customers may enter addresses in Arabic, geocoders may expect Western address formats, and driver apps may display English.	Each translation layer can introduce data loss, reducing first-attempt delivery and ETA accuracy.

This is why the question is not only whether a routing engine can sequence stops. It is whether the system can trust the stop location in the first place. Poor geocoding accuracy, bilingual inputs and address ambiguity all reduce the quality of the route before the optimiser begins its work.

The financial consequence compounds quickly. A failed delivery attempt adds driver time, fuel, re-dispatch cost, customer-service cost and usually a WISMO — “where is my order?” — interaction. It also damages customer experience and lifetime value. In high-volume networks, failed deliveries are not just exceptions; they are recurring margin leakage.

According to Kearney Middle East, the region’s e-commerce market has continued to grow at double-digit rates, outpacing most mature markets. That growth makes the addressing problem more expensive with every order added to the network. For enterprise retailers, marketplaces, grocers and 3PLs scaling in the GCC, failed delivery rate is not just an operations KPI. It is a growth constraint.

For Locus, this is where last-mile optimisation starts: not with the best theoretical route, but with a reliable delivery point that dispatchers, drivers, customers and carriers can all act on.

The Four-Layer Architecture: What AI Route Optimisation Actually Requires in the GCC

The platforms that produce real cost reduction in the region do not treat route optimisation as a single feature. They treat it as an operating architecture with four connected layers, each designed for GCC conditions.

Layer 1: Location Intelligence — The Foundation

This is where many global routing platforms underperform in the region. A GCC-ready location intelligence layer must do four things well:

• Treat local address systems — Makani, Saudi National Address and Qatar zones — as first-class inputs, not fallback fields.
• Parse Arabic and English addresses natively, so the same address entered in either language resolves to the same delivery coordinate.
• Use reverse geocoding with landmarks, points of interest and customer phone-location signals when the written address is ambiguous.
• Learn from completed deliveries. When a driver taps “delivered here” at a coordinate that differs from the original geocode, that refinement should be stored and applied to future orders at the same address.

If the location layer is weak, the optimisation layer is forced to solve the wrong problem.

A Riyadh-based e-commerce brand experiencing geocode mismatches across new residential compounds does not solve that problem with a more sophisticated route sequence. It solves it by improving address confidence before dispatch. Every confirmed delivery point should strengthen the address database, reduce future exceptions and improve SLA adherence.

Layer 2: Route Optimisation Engine

Once every order has a reliable coordinate, the optimisation engine can do meaningful work. A GCC-native route optimisation engine must evaluate several constraints at the same time:

• Heat: driver welfare, shift design, vehicle cooling capacity and exposure windows during peak summer months in the UAE and Saudi Arabia.
• Traffic patterns: Sheikh Zayed Road peak periods, Dubai Metro-linked congestion cycles, Riyadh ring-road flows and Doha arterial congestion.
• Prayer time windows: five daily windows that shift by season and affect customer availability, driver scheduling and service expectations.
• Ramadan windows: Iftar and Suhoor patterns that compress delivery demand into narrower operational bands, requiring teams to optimise deliveries during Ramadan peak periods.
• Customer time preferences: including afternoon and post-Asr delivery demand in many GCC markets.
• Multi-drop density versus distance trade-offs: balancing route clustering, promised delivery windows, vehicle capacity and driver hours.

The architectural distinction matters. Generic route optimisers often create a route and then filter it against local realities. GCC-native systems optimise with those realities built in from the start. That difference shows up in fewer kilometres per drop, higher route density, better on-time delivery and lower cost-to-serve.

Layer 3: Dispatch and Execution

The GCC courier market is fragmented. Regional operators such as Aramex and SMSA operate alongside national postal operators, city-level 3PLs, marketplace fleets, gig capacity and brand-owned vehicles.

Dispatch automation must therefore orchestrate across multiple capacity pools, not simply assign routes to a single fleet. Intelligent dispatch should select the right fleet or carrier by lane, cost, capacity, service commitment and live performance — not by static contract rules alone.

Execution also requires real-time re-routing. If a driver corrects an address at the doorstep, that correction should influence live routes for other drivers in the same area where relevant. It should not sit unused until the next quarterly reporting cycle.

For enterprise operators, this is where AI route optimisation becomes an execution system: dispatchers move from manual exception handling to automated decision control, with the system continuously balancing SLA adherence, delivery density and carrier cost.

Need a GCC-ready last-mile transformation plan?

Work with supply chain experts to assess address quality, dispatch workflows and routing performance across UAE, Saudi Arabia and Qatar.

Talk to an Expert ?

Layer 4: Learning and Feedback

Every delivery should improve three models:

1. The geocoding model — improving coordinate accuracy and address confidence.
2. The ETA model — learning from local traffic, heat, access delays, prayer windows and dwell time.
3. The failed-delivery prediction model — identifying orders likely to fail before a driver leaves the hub.

The system becomes more accurate, and more profitable, the longer it runs in the region. That feedback loop is critical in fast-growing markets where new districts, compounds and fulfilment nodes are constantly being added.

According to McKinsey & Company, AI and advanced analytics are among the highest-impact capabilities for supply chain performance, particularly in markets where logistics data and infrastructure are maturing in parallel. That describes the GCC well: strong logistics infrastructure, high service expectations and a growing need for operational intelligence at the last mile.

Route Optimisation AI Workflow for Middle East Logistics

A GCC-ready AI route optimisation workflow should operate as a closed loop rather than a one-time planning exercise.

Workflow stage	What happens	GCC-specific requirement
1. Data ingestion	Orders, addresses, promised windows, vehicle data, driver shifts, carrier capacity and traffic inputs are collected.	Inputs must support Arabic and English address formats, Makani, Saudi National Address, Qatar zones and customer time preferences.
2. Address normalisation	The system cleans, parses and validates delivery locations before dispatch.	Descriptive addresses, landmarks, compound gates and bilingual variants must resolve to reliable coordinates.
3. Constraint modelling	Business rules and real-world limits are converted into planning constraints.	Heat, prayer windows, Ramadan demand patterns, driver welfare and city-specific traffic must be treated as first-class constraints.
4. Optimisation	The route engine sequences stops, assigns vehicles and balances cost, density and SLA risk.	The optimiser must avoid solving for shortest distance alone; it must solve for executable delivery plans.
5. Dispatch execution	Routes are assigned to drivers, owned fleets, 3PLs or carriers.	Carrier orchestration must account for fragmented GCC capacity pools and live performance.
6. Real-time adjustment	Routes are re-sequenced as traffic, address corrections, cancellations or driver delays occur.	Dynamic re-routing must be fast enough to protect delivery promises during peak city congestion.
7. Feedback loop	Completed deliveries, failed attempts, dwell time and driver confirmations feed back into future models.	The system should improve geocoding, ETA accuracy and failure prediction with every delivery cycle.

This workflow is what separates route planning software from an AI-powered last-mile operating system.

Five Ways AI Route Optimisation Cuts Last-Mile Costs in the GCC

For enterprise e-commerce, grocery, retail and 3PL operations across the UAE, Saudi Arabia and Qatar, five specific levers drive most of the cost impact.

Cost lever	Operational mechanism	KPI impact
Address disambiguation	Resolves vague, bilingual or landmark-based addresses before dispatch	Higher first-attempt delivery, fewer re-dispatches
Constraint-based routing	Optimises around heat, prayer windows, traffic and customer preferences together	Lower kilometres per drop, higher route density
Dynamic ETA accuracy	Learns local traffic, dwell time and customer availability patterns	Fewer WISMO calls, better on-time delivery
Carrier orchestration	Allocates orders across owned fleet, 3PLs and carriers based on cost, capacity and SLA	Lower cost-to-serve, stronger SLA adherence
Predictive failure detection	Flags risky deliveries before dispatch and triggers confirmation workflows	Fewer failed attempts, less driver idle time

1. Address Disambiguation Reduces Failed First-Attempt Deliveries

This is the largest cost lever in many GCC e-commerce operations. A retailer improving first-attempt delivery by resolving ambiguous locations removes avoidable re-dispatch cost from every 1,000 orders. Fuel, driver hours, WISMO call volume and customer-experience damage all move in the right direction.

The mechanism is not simply a better route. It is bilingual geocoding, landmark-based reverse lookup and customer phone-location signals working before dispatch.

2. Constraint-Based Routing Cuts Kilometres Per Drop

Route density improves when heat, prayer times, traffic and customer preferences are optimised together rather than sequentially.

A Dubai operator clustering deliveries around post-Asr windows can protect customer convenience while reducing dead kilometres and vehicle underutilisation. In peak summer months, that can affect not only fuel cost but also fleet planning and driver shift design.

3. Dynamic ETA Accuracy Reduces WISMO and Driver Idle Time

GCC customers do not want a delivery promise that says “between 9am and 6pm”. They expect narrower windows and proactive communication.

ETA models trained on local conditions — Sheikh Zayed Road, Riyadh ring roads, Doha arterial congestion, building access delays and prayer-time availability — support more credible promises. The operational benefit is straightforward: fewer WISMO calls, less dispatcher intervention, fewer failed handovers and better SLA adherence.

4. Carrier Orchestration Across Fragmented GCC Networks

The region’s courier market spans Aramex, SMSA, national postal operators, city-level 3PLs and owned fleets.

AI-driven carrier allocation should select the lowest total-cost service option per shipment based on live lane performance, capacity, SLA risk and cost — not only annual contract rules. According to PwC Middle East, consumer and retail transformation in the GCC is being driven by customer-experience expectations that outpace legacy logistics infrastructure. In that context, carrier orchestration is a margin lever, not an optional feature.

5. Predictive Failure Detection Prevents Failed Attempts Before They Happen

Machine learning models can flag high-risk deliveries before dispatch: vague addresses, historically unreachable coordinates, customer time-window conflicts, or delivery slots that collide with known prayer-time or Ramadan behaviour.

The system can then trigger a customer confirmation workflow before the driver leaves the hub. That converts a likely failed attempt into a preventable exception.

According to the Saudi Vision 2030 national strategy, logistics is one of the pillar sectors targeted for transformation. That raises expectations for technology-enabled last-mile performance in the Kingdom and, through competitive pressure, across the wider GCC.

AI Route Optimisation vs Traditional Route Planning

Traditional route planning and AI route optimisation are often discussed as if they solve the same problem. They do not.

Traditional planning usually depends on static rules: fixed territories, dispatcher experience, historical route templates and map-based distance calculations. It can work when demand is predictable and exceptions are low. GCC last-mile operations rarely meet those conditions.

AI route optimisation uses operational data to continuously adjust delivery decisions as the day changes. It can re-sequence stops, rebalance driver loads, account for real-time traffic, flag likely failures and recommend the best fleet or carrier for each shipment.

Capability	Traditional route planning	AI route optimisation
Route design	Static or dispatcher-led	Dynamic and constraint-based
Address handling	Manual correction after failure	Address confidence before dispatch
Traffic response	Limited or reactive	Real-time re-optimisation
Customer time windows	Often treated as fixed rules	Optimised with capacity, density and SLA risk
Carrier allocation	Contract-led or manual	Data-led across owned fleets, 3PLs and carriers
Learning loop	Reports issues after the fact	Improves future routes through field feedback

For Middle East operators, the distinction matters because the operating environment changes quickly. Traffic, customer availability, building access, district growth and delivery density are not static variables. A routing system that cannot learn from those conditions will keep producing routes that look efficient on paper and fail in execution.

Manual vs AI Route Planning in GCC Operations

Operating metric	Manual or static planning	AI route optimisation
Planning time	Dispatcher-led, often hours for larger fleets	Automated sequencing, with reported reductions from 2–3 hours per day to 5–10 minutes for 50 trucks
Fuel efficiency	Dependent on dispatcher experience and static assumptions	AI-enabled routing and dispatching can deliver 10–15% fuel savings
Distance per delivery	Often inflated by low route density and poor sequencing	Reported reductions of 12–20% fewer kilometres per delivery
On-time delivery	Vulnerable to traffic, address corrections and capacity changes	Reported improvement from 75–82% to 92–97% in GCC fleets adopting AI route optimisation
Local constraints	Often handled manually after route creation	Heat, prayer windows, Ramadan and customer preferences can be embedded into the optimisation logic
Continuous improvement	Limited to periodic reporting	Learns from geocoding corrections, dwell time, failed attempts and driver feedback

Benefits of AI Route Optimisation for Middle East Logistics Teams

AI route optimisation creates value across cost, service and control. The highest-impact benefits for GCC operators include:

Lower Cost-to-Serve

Better address accuracy, denser routes and automated carrier selection reduce avoidable kilometres, re-dispatches, driver idle time and manual dispatcher work.

Higher First-Attempt Delivery

When the system validates delivery points before dispatch and learns from completed deliveries, fewer orders fail because of vague addresses, wrong gates or unreachable coordinates.

Better On-Time Delivery

Dynamic routing and ETA models trained on local traffic, dwell time and customer availability improve promised-window adherence.

Stronger Fleet Utilisation

AI can balance volume across owned vehicles, 3PL partners and gig capacity, helping operators reduce underutilised assets while protecting SLA performance.

Reduced WISMO Volume

More accurate ETAs and proactive communication reduce customer-service burden. Customers do not need to ask where the order is if the system can provide credible, real-time visibility.

Better Dispatcher Productivity

Dispatch teams move from manual planning and exception firefighting to exception supervision, performance control and continuous improvement.

More Resilient Operations

AI-supported routing can respond faster to traffic disruption, address corrections, vehicle availability changes and demand spikes during Ramadan, sales periods or seasonal peaks.

Lower Emissions Exposure

Fuel savings and reduced kilometres per delivery also support sustainability goals. When routes are denser and idling is reduced, fleets can lower emissions intensity per order while improving operating economics.

Key Features to Look for in AI Route Optimisation Software for the Middle East

Enterprise buyers in the GCC should evaluate route optimisation platforms against the realities of regional execution, not generic software claims.

1. GCC-Ready Geocoding

The platform should support Makani, Saudi National Address, Qatar zones, landmarks, POIs and historical delivery confirmations.

2. Arabic-English Address Parsing

Bilingual address handling should be native to the platform. If the system depends on dispatcher interpretation or customer-side transliteration, it will introduce operational drag.

3. Dynamic VRP Optimisation

The engine should handle vehicle capacity, driver hours, stop priority, promised windows, service time, route density and cost constraints together.

4. Prayer-Time, Ramadan and Heat Constraints

Regional operating conditions should be built into route planning, not added after the route is already created.

5. Real-Time Re-Routing

The system should update routes when addresses are corrected, customers become unavailable, vehicles are delayed or traffic patterns change.

6. Multi-Fleet and Carrier Orchestration

GCC operators often use a mix of owned fleets, 3PLs, national carriers and city-level partners. The platform should allocate work dynamically across that network.

7. Driver App and Field Feedback

Drivers should be able to confirm locations, report exceptions, capture proof of delivery and feed operational data back into the system.

8. KPI-Level Reporting

The platform should measure business outcomes such as first-attempt delivery, cost per drop, kilometres per order, SLA adherence, driver utilisation, WISMO volume and carrier performance.

9. TMS, ERP and Telematics Integration

AI routing must connect to the systems that run the operation: TMS, ERP, WMS, order management, telematics, driver apps and customer communication workflows. Without integration, optimisation remains a planning layer rather than an execution capability.

The VP’s Evaluation Framework: Four Questions Before Investing

Before approving the next last-mile technology investment, supply chain leaders in the region should pressure-test the platform against four questions:

1. Does the platform have a GCC-native location intelligence layer — Makani, Saudi National Address and Qatar zones — or is it retrofitting a global geocoder?
2. Can it parse Arabic and English addresses natively, or does it rely on customer-side transliteration?
3. Does its optimisation engine treat prayer windows, heat, Ramadan and GCC traffic patterns as first-class constraints, or as afterthought filters?
4. Does it orchestrate dynamically across fragmented GCC carrier networks, or does it assume a single-fleet operating model?

For buyers comparing vendors, the evaluation should also include implementation maturity, data readiness and how to choose route planning software that can support the region’s operating realities.

A useful way to evaluate the difference:

Evaluation area	Retrofitted global tool	GCC-native AI route optimisation
Address handling	Relies heavily on standard geocoding and manual correction	Learns from Makani, Saudi National Address, Qatar zones, landmarks and delivery confirmations
Language support	Often depends on transliteration or dispatcher review	Parses Arabic and English inputs as operational data
Local constraints	Applies heat, prayer and Ramadan rules after routing	Optimises routes with these constraints built in
Dispatch model	Assumes single fleet or static carrier allocation	Automates allocation across owned fleet, 3PLs and carriers
Performance learning	Reports exceptions after the fact	Feeds address, ETA and failure data back into future plans

If any of the four answers is “no” — or “partially” — the economics of the investment will underperform the business case. Not because the algorithm is inherently wrong, but because the operating model it was designed for is not the GCC.

Why Choose Locus for AI Route Optimisation in the Middle East?

Locus is built for real-world logistics execution, where routing is only one part of the operating problem.

For GCC enterprises, Locus helps connect the full last-mile workflow:

• Location intelligence to improve address confidence before dispatch
• Route optimisation to sequence stops around real constraints
• Dispatch automation to allocate work across fleets, carriers and capacity pools
• Real-time visibility to monitor execution and respond to exceptions
• ETA intelligence to improve customer communication and reduce WISMO
• Feedback loops to improve address, route and failure-prediction models over time

The point is not to produce the mathematically shortest route in isolation. The point is to produce the most executable plan for the region’s actual delivery environment — one that improves route density, cost per drop, SLA adherence and on-time delivery together.

For Middle East logistics leaders, the platform question is no longer simply whether software can generate routes. It is whether the system can interpret local delivery points, model regional constraints, automate dispatch decisions and improve with every delivery.

Scale high-volume delivery execution beyond spreadsheets

See how automated routing and dispatch improve SLA adherence, fleet utilisation and delivery control across complex distribution networks.

See the Use Case ?

The Real Question GCC VPs Should Be Asking

The winning e-commerce, retail, grocery and 3PL operations across the UAE, Saudi Arabia and Qatar over the next five years will not simply be those with the most advanced routing algorithms. They will be those whose last-mile technology stack is engineered for the GCC from the data layer up.

That means the system must understand how local addresses are written, how customers choose delivery windows, how heat and prayer times affect execution, and how the regional carrier market is structured.

The question for supply chain leaders is not, “Should we invest in AI route optimisation?”

It is: is our last-mile technology stack designed for the GCC — or imported from somewhere else and hoping to fit?

For Locus, that is the operating principle behind effective AI route optimisation in the Middle East: start with location confidence, optimise against real-world constraints, automate dispatch decisions, and improve every route through feedback from the field.