09 June, 2026

AI for logistics costs: cutting 15% of wasted delivery mileage

Quick summary: AI cuts logistics costs mainly by re-planning delivery routes: the computer evaluates millions of options in minutes and picks the shortest route that still meets real constraints like truck-ban hours, vehicle load and customer receiving windows. The waste left by manual route planning usually falls around 15 to 22% of fleet operating costs, based on the surveys we ran for mid-sized distribution businesses. This article explains where the money goes, how far routing software can take you, and when it isn’t worth the investment.

Transport costs spike because routes are planned by gut feeling

Every month, the operating cost sheet gets thicker. You can clearly see fuel, tolls and driver wages eating into profit. But there is one cost that drains out quietly: the extra distance nobody records in any ledger.

For a business with 30 trucks, that excess can reach hundreds of millions of VND a month. It is real money, it just doesn’t sit on its own cost line, so it is very hard to see.

The hidden cost of planning routes on experience

You trust your dispatch team, and they deserve it, they are veterans. But even the best person is only human. No one can compute every variable in real time: traffic jams, hourly truck bans, bridge weight limits, the delivery times promised to customers.

Every decision based on experience carries risk. A few extra kilometers of detour. Fifteen minutes waiting at an intersection. They add up, and become a hidden cost nobody notices until the end of the reporting period.

Overlapping delivery route map driving route costs up
A few extra kilometers per trip, multiplied across thousands of trips, is a very different number.

Fuel and vehicle wear: the next cost that gets overlooked

Waste doesn’t stop at extra distance. It is fuel burned for nothing every day, tires wearing faster than expected, maintenance arriving ahead of schedule, and driver hours stretched without need.

We once worked with a consumer goods distribution partner with 28 branches and a fleet of 80 trucks. Their problem was exactly this: everything ran, it just ran expensively. After replacing manual routing with an automated route planning system, the system calculated thousands of options by itself and picked the optimal route from traffic data, cost and real-time conditions. In the first quarter they cut 13% of fleet operating costs, worth over 580 million VND kept.

A busy dispatch room, yet fleet capacity stays low

Every month you sign off the dispatch team’s payroll, and it is no small figure. Yet what comes back is half-empty trucks returning, routes crossing over each other, and fuel and depreciation costs climbing.

That is the dull, constant pain: high pay for low productivity. Dispatchers look very busy, hunched over screens all day, yet the whole team’s capacity doesn’t budge.

A dispatch room full of route notes pinned up, with the MONA bear standing next to a staff member on the phone
The dispatch team is busy all morning, yet fleet capacity doesn’t move up.

The limits of a human planning thousands of delivery points

The problem isn’t diligence, it is the natural limit of the brain. Even the most experienced dispatcher can’t plan 500 delivery points optimally at once, let alone 2,000 points a day with dozens of attached constraints. They have to rely on gut feeling and old experience, and mistakes are inevitable.

In the surveys we ran for mid-sized logistics businesses, the waste from manual route planning usually falls around 15 to 22% of fleet operating costs. The figure varies with delivery density and constraint complexity, but we haven’t seen a case where it dropped below 10%.

The knock-on effect on the customer side

Non-optimal routes lead to a very predictable chain of effects. Drivers go farther. Goods arrive later. Customers complain. Then you lose customers and your reputation, all because of one dispatch step.

For a supply chain with 42 warehouses in southern Vietnam, we deployed an automated route-assignment module inside the transport management system. The system processes 18,000 orders a day, merging orders and distributing the load across 250 trucks. Average delivery time per trip dropped by 35 minutes, and on-time delivery rose from 89% to 97.8% after 3 months.

AI cuts logistics costs with custom-built routing software

The market talks too much about AI. Generative AI, analytical AI, predictive AI. Every operations director we meet has heard those promises, tried a few off-the-shelf tools, and found they don’t touch their real problem.

Your problem is far more specific: how do 120 trucks deliver to 1,800 drop-off points across 15 provinces every day without wasting a liter of fuel, and how do you keep drivers from circling for 2 hours just because the navigation system doesn’t know that alley is closed to trucks?

One consumer goods distribution chain was burning over 800 million VND a month on excess operating costs. They had 5 dispatch staff who spent 4 hours every morning just planning routes by hand. To them, the word AI sounded both distant and risky, and we think that reaction is reasonable.

So we don’t sell the word AI. We write business process automation software, where route calculation is just one module inside, serving the business’s exact problem.

MONA software showing re-planned routes on a multi-province map
The machine runs through every option before choosing, while a person only has time to pick the first option that sounds reasonable.

Encode every real constraint into the software

Every business operates in its own way, so we don’t apply a generic formula. The first job is to sit down with the dispatch team and write out every real constraint, down to dry detail. For example, the system must know:

  • Truck-ban hours on routes in Ho Chi Minh City and Hanoi.
  • The maximum load of each vehicle type, from 1.9 tons and 3.5 tons up to 8 tons.
  • Receiving windows for over 2,000 wholesale customers, each with its own pattern.
  • Special conditions such as delivery points that need a refrigerated truck, or drivers who must hold a separate safety certificate.

Only after loading all 74 of that business’s specific constraints does the routing module start running millions of scenarios. The result is a plan that balances both cost and time, not simply the shortest road on the map.

For the distribution chain above, the system finished planning routes for 85 trucks to 1,200 delivery points in under 15 minutes, replacing work that used to take 5 people half a morning. The whole fleet’s total distance dropped 18% in the first month.

Connect to the warehouse and order systems already running

Good routing software can’t stand alone, it has to be a link in the chain. We understand you have already invested in a management system, customer management or warehouse management software in the past, so our job is to connect them, not make you replace everything.

The approach is to build dedicated connection gateways so order data from the sales system flows into the routing software on its own, with no manual entry. Removing manual entry alone eliminates most errors like wrong addresses or wrong quantities.

Once routes are planned, the information is pushed automatically to the driver’s phone app and to the warehouse system. The warehouse knows exactly which order to pick for which truck, and drivers know the delivery order and contact details for each stop. No more calling each other on the road.

Four numbers you can measure this week

Before you think about investing in software, get four numbers from your own fleet. One: the total kilometers the whole fleet drove last month. Two: the average load factor per trip, meaning how full the truck is compared with its capacity. Three: how many hours the dispatch team spends each morning planning routes. Four: last month’s on-time delivery rate.

Those four numbers come from vehicle logs and order tracking sheets, no software needed. With them, any pitch about route optimization can be checked right away, because you know where you stand. In the projects we have run, the biggest gain usually sits in the load factor, not the distance.

A 25-branch distribution chain keeps 1.2 billion VND a month

Fuel costs inflate for no clear reason. Drivers report traffic jams, but the GPS shows an unfamiliar road. Goods arrive late. Customers complain. That feeling of losing control over costs is something many distribution business owners face every day.

We understand that feeling, and it isn’t only about money. It is the helplessness of having blind spots in your own operation that you can’t measure. A fast-moving consumer goods distribution chain with 25 branches fixed exactly that with an automated route planning system.

Fuel and maintenance down 18%

Fuel and vehicle wear costs are a number that speaks, reflecting delivery team efficiency directly. Before, their drivers chose routes on their own experience, which led to many unnecessarily long runs and a lot of empty driving time.

After switching to automated routing based on real traffic data, the number of delivery points and vehicle load, fuel costs fell 18%. Trucks ran less, so maintenance came less often too. This line alone kept over 550 million VND a month.

Peak-hour delivery capacity up 25%

Peak hours are when things break most easily. One late trip sets off a chain reaction: dispatch tension, impatient customers, revenue at risk.

The system doesn’t only plan the route, it also assigns orders to the driver best positioned. As a result, with the same fleet and the same drivers, delivery capacity between 9 and 11 a.m. and between 2 and 4 p.m. rose 25%. Goods reach customers faster and on time.

This routing approach can be reused across many other industries

Many of you think route optimization is only for container trucks and large trucks. That thinking is understandable but no longer true. Any business with staff moving around outside faces exactly this problem, so how AI cuts logistics costs for a truck fleet applies just as well elsewhere.

Many different industries share the same route optimization problem
Any industry with people on the road has the same routing problem.

Installation and technical maintenance services

How much time does your technical team spend every day just traveling? A customer in District 2 waits from morning to afternoon, while another technician drives from District 9 back to Go Vap, passing right through District 2. It is wasteful, burns fuel and costs you credibility.

For an electronics retail chain with 32 stores, we built self-dispatching software: the system groups appointments in the same area and draws an optimal route for each technician every morning. They cut 22% of excess distance and raised daily customer support visits by 15% without adding people.

Pharmacy chains and mini-supermarkets

Small trucks deliver to dozens of branches every day, and the route depends entirely on the driver. Sometimes they pass through District 7 and only then remember a drop in District 4, so they have to turn back. Goods arrive late, branches are left reacting, and fuel money keeps draining.

At a system of 45 mini-supermarkets, the dispatch software was connected directly to the warehouse management system, scheduling trips and delivery routes automatically from the branches’ orders. Internal delivery time dropped 30%, and they kept over 300 million VND a month from this fleet optimization alone.

Waste collection and environmental services

This industry has its own quirk: trucks run on very rigid fixed routes. Many collection points aren’t full yet but the truck still has to arrive on schedule, while other areas with sudden needs go unnoticed. Labor and fuel get poured into unnecessary trips.

The fix is to fit fill-level sensors to industrial bins, push the data to a central system, then let the software re-plan collection routes every day, prioritizing points that are full or nearly full. That alone cut 18% of truck trips.

Express delivery and food delivery

Customer care for business clients that comes with orders is a separate branch of work, see AI agents replacing B2B customer service.

In this industry speed is everything, competing minute by minute. Drivers take several orders at once and must estimate routes themselves, which leads to late deliveries, cold food and cancelled orders. Every cancelled order costs money and a customer.

A real-time order batching and route recalculation module handles exactly this: the system groups nearby orders for one driver and keeps updating the route with traffic conditions. One food delivery partner cut average delivery time per order by 25%, and the cancellation rate from late delivery fell from 7% to under 2%.

When routing software isn’t worth the investment yet

We say this part plainly so you don’t lose money for nothing, because most money lost for nothing in automation comes from buying earlier than you need. We covered those kinds of losses in the article 5 mistakes of cheap automation. There are three cases where routing software isn’t worth building yet.

One: a fleet of fewer than 8 to 10 vehicles and fewer than 60 delivery points a day. At that scale, a skilled dispatcher still plans close to optimal, and the extra gain from software doesn’t cover its cost.

Two: customer address data is still too dirty. If more than a third of the addresses in your system aren’t standardized or don’t have coordinates, the job to do first is cleaning data, not buying software. Routing on wrong addresses gives wrong results.

Third is when the delivery route is nearly fixed and rarely changes, for example exactly 12 stops every day, in the same order, in the same time window. In that case you only need to standardize the schedule once, with no need for a dynamic routing system.

AI cuts logistics costs: what you can do this week

Look back at the numbers in this article: 15 to 22% of operating costs go to manual route planning, 5 people spend 4 hours every morning splitting routes, over 800 million VND per month in excess cost at one distribution chain, and 550 million VND saved per month on fuel alone at a 25-branch chain.

We know you have already put effort into finding a fix: adding people to the dispatch room, buying GPS devices for the trucks, making drivers report routes before each morning, paying by the kilometer, even outsourcing part of the fleet. What matters is that you went looking instead of just putting up with it. But all of those fixes hit the same wall: people still have to plan the routes by hand, and the number of constraints is more than one head can hold in a single morning.

Where AI that cuts logistics costs clearly beats people is this: it runs through every option before choosing, while a person only has time to pick the first option that sounds reasonable. What remains is encoding your company's constraints correctly into the software, and that part has to be done sitting together with your dispatch team.

Bring four fleet numbers, and we will run a trial calculation in 60 minutesTotal kilometers the whole fleet drove last month, average load rate per trip, hours the dispatch room spends every morning splitting routes, and on-time delivery rate last month. Those four numbers from your vehicle logs and order tracking sheet are enough. You show them to us, and we sit down and work out what a machine can carry, what is best left alone because a machine would do no better than a person, and roughly how much the savings come to per month. MONA does not charge for that session, and if your fleet is not yet big enough to justify route-planning software, we will say so plainly in the first session. MONA has built software since 2016, has handled over 14,000 projects and keeps an 85% returning-client rate, so we have no need to sell you a project that is not worth doing. Call 1900 636 648, and our consulting team will call back within a working day.

Frequently asked questions about AI cutting logistics costs

How much can AI cut logistics costs?
In the deployments we have run, the savings usually land at 13 to 18% of fleet operating costs in the first quarter. For a chain of 28 branches with 80 trucks, a 13% reduction is worth over 580 million VND a month. For a chain of 25 branches, an 18% fuel reduction alone kept over 550 million VND a month.
Why does manual route planning waste up to 15% of costs?
Because a dispatcher can’t optimize hundreds of constraints at once: traffic jams, hourly truck bans, bridge load limits, each customer’s delivery window. Every small deviation, a few extra kilometers of detour or fifteen minutes of waiting, adds up across thousands of trips into a hidden cost of hundreds of millions of VND a month.
How many vehicles should a business have before investing in routing software?
The practical threshold we usually use is around 10 vehicles and over 60 delivery points a day. Below that, a skilled dispatcher still plans close to optimal, and the extra gain from software doesn’t cover its cost.
Can routing software connect to the warehouse and sales systems we already use?
Yes, and it is a requirement, not an option. The approach is to build a connection gateway so orders from the sales system flow straight into the routing software, then push the planned routes back down to the driver app and to the warehouse system so goods are picked for the right truck.
What data do we need to prepare before deployment?
Three groups of data: a standardized customer address list with coordinates, fleet information including load capacity and body type, and delivery constraints such as receiving windows and truck-ban hours by route. If more than a third of addresses aren’t standardized, clean the data first, because routing on wrong addresses gives wrong results.

Main topic: AI in logistics

About The MONA Group: Founded in 2016, with over 14,000 projects delivered and an 85% client retention rate. MONA does website design, full-scope SEO, custom software and automation software for businesses. Consulting hotline: 1900 636 648. See also MONA AI Agent suite and automation software development.
MONA has a podcast: listen instead of readingEpisodes on AI, automation & SEO for Vietnamese businesses
Nghe ngay
About the author

CEO & Founder · The MONA

The founder of The MONA built a service company of 60+ staff with 14,000+ projects delivered. Khánh Hùng shares his thinking on building self-running systems, automating a business, and the real lessons behind MONA's projects.

Views 400
Rate this article 4,8/5 · 15 ratings

Professional website
design services

Own a website with a beautiful, 100% exclusive design and a lifetime warranty, with the ability
to expand features flexibly as your business grows, starting today!

Contact Mona

    MONA có riêng một Người "Bạn Thân" cho bạn - Người Account sẽ đồng hành, hỗ trợ, hướng dẫn, đặt đồ ăn cho bạn mãi mãi, từ đây về sau!
    MONA will never sell your information or send you spam
    The MONA team advising on an online education management system
    Hỏi đáp giáo dục 4.0
    Tạo cuộc hẹn miễn phí với MONA để giải đáp và tư vấn mọi thắc mắc về giải pháp số hoá ngành giáo dục
    Thời lượng cuộc hẹn
    45 Phút
    Ngày và giờ
    Thứ 2, ngày 25 tháng 12, 2023
    [9:30 - 10:15]

      Chọn ngày và giờ
      Khung giờ
      Back
      Hãy cho MONA biết bạn là ai
      0:00