40S
← Profiles
01
↓ / → to advance
The operating system for logistics

Logistics that
runs itself.

40S is an AI-native control layer that sits on a live logistics operation, predicts the failures before they happen, and acts on them autonomously. In English and in Arabic.

30M+
orders orchestrated
28
production AI agents
14
live ML models
24/7
autonomous voice · EN + AR
Investor brief 2026 40s.ai
The problem

Logistics drowns in
coordination work.

Every delivery generates a swarm of small tasks: calling the customer who is not home, chasing the driver on WhatsApp, reconciling a cash-on-delivery amount, categorizing a failed delivery, filing a customs declaration, answering "where is my order".

Each task is tiny. In aggregate they consume the majority of a 3PL's operational payroll and put a hard ceiling on how fast it can grow.

And the tools meant to help are reactive. A dashboard shows you what already broke. By the time a human reads it, the delivery has failed, the customer is angry, and the cash is heading back to the warehouse.

The reactive loop, today
Something breaks
↓
A human notices, eventually
↓
They ask the AI a question
↓
They act, by hand, too late

Every logistics company on earth has this loop. It is a customer-service pattern, not an operating system.

Why it matters

The last mile is where
the money leaks.

Share of total shipping cost
53% last mile 47% everything else Statista, 2023. Up from 41% in 2018.

It is also the most failure-prone leg. A failed first attempt costs two to three times the original delivery to redo. In cash-on-delivery markets, a third failure sends the order, and the cash, back.

~30%
of all support contacts are "where is my order", at $4 to $12 each. BenchmarkPortal, multi-source
2 to 3x
the cost to re-attempt a failed delivery vs. getting it right the first time.
~30%
of actual UAE orders are still cash-on-delivery, carrying return and cash-handling risk. Shorages, 2024
>40%
of operating expense is labor for the majority of 3PLs. Extensiv, 2024
Why now

Three things moved
at once.

01

Models crossed the line

Frontier models became reliable enough for customer-facing voice and sustained multi-step autonomy. The bottleneck moved from intelligence to the surrounding infrastructure: data, tools, telephony, permissions, escalation. That is exactly where 40S is built.

02

Arabic became first-class

Voice AI now handles Gulf Arabic with sub-100ms speech latency. A product built here, for here, no longer carries a language penalty against an English-first product retrofitted for the region.

03

The region is investing

GCC e-commerce keeps compounding and 3PLs need to scale without adding headcount. Saudi Vision 2030 alone commits roughly $133B to logistics. The agentic-AI field in the region is still contested by a handful of players.

The window to build the category-defining local brand is open right now.

Act II · The product

From watching
to acting.

Dashboards tell you what already broke. 40S sees it coming and does something about it. Here is how.

The product

One substrate.
Skills on top.

A carrier that adopts 40S does not buy a dispatch tool, and a tracking bot, and a finance package, and a call center. It buys one runtime where orders, drivers, merchants, finance, and customer communication all flow through the same agents, on the same live data.

This is the bet Microsoft made with Windows and Salesforce made with the platform: own the substrate, and every new capability is a skill added to it, not a new vendor.

Tracking and WISMONDR recoveryFinance and AP/AR Customs (Bayan)VoiceReportingOnboarding

The agent runtime

28 specialist agents, 8 to 20 tools each, on an approval-gated write system with full before-and-after state and audit trail.

Architecture

From live data to
autonomous action.

Built AI-first on top of any production operation, multi-carrier, not bolted onto a legacy stack.

live

Production data

Carrier systems sync every 30 seconds into an operational store plus a warehouse.

predict

14 ML models

Failure, ETA, returns, demand, driver match, churn. Re-scored continuously.

decide

Intervention engine

Scans for risk, opens an intervention, picks the recommended action.

act

28 agents

Execute inside guardrails. Low risk runs autonomously, high risk asks first.

reach

Omnichannel

Voice, WhatsApp, SMS, email. Arabic and English, natively.

The guarantee

Agents can analyze anything freely. Anything that mutates a live system is role-scoped, auditable, reversible, and risk-tiered. That is what makes autonomy safe enough for a real operation.

The shift

From answering questions
to preventing failures.

40s.ai / dashboard
Autonomous Intelligence
7 interventions today
18
At-risk orders
6 stuck · 7 SLA · 5 multi
23
Failed + returned
14 failed · 9 returned
AED 84k
COD at risk
AED 312k saved (30d)
High-risk order likely to fail
failure_prediction_v1 scored 0.92 · AED 2,500 COD · 2 attempts · recipient unreachable
~AED 2,500 · call customer ›
SLA breach imminent
ETA passed for 3 Business Bay orders · driver behind route
~AED 640 · notify + re-sequence ›
Return likely, COD at risk
return_prediction_v1 0.88 · merchant return rate elevated
~AED 1,180 · proactive WhatsApp ›
Live dashboard · the intervention feed, acting on real orders
The new loop

AI detects the risk hours early. It reassigns the driver and calls the customer with a new window. The delivery succeeds. The operator sees a resolved problem in the morning briefing.

L1 Visibility L2 Intelligence L3 Augmentation L4 Supervised autonomy · now L5 Autonomous ops

Gartner expects 60% of supply-chain disruptions resolved without human intervention by 2031. That is precisely this engine. Gartner, 2026

The call

And it makes the call itself.

A real outbound call, in the customer's language, transcribed and scored live, then resolved. This is the autonomous action from the previous slide, happening on the phone.

R
Rami Al Hendi
Live·00:00·Outbound · Arabic
Live analysis
Concerned
Talk ratio
Agent 46%Customer 54%
Detected on the call
Order 40SDEMO001 COD AED 2,500 New window: this evening
Outcome → Reschedule confirmed, delivery saved, COD protected. Logged and fully reversible.
Perception

40S listens to every call.

Every customer call is transcribed, scored for sentiment, classified, and checked for resolution, turning a call center into a real-time signal feed. Numbers below are from a single live client deployment.

0
calls analyzed, one client
0
resolved on the call
0
flagged negative sentiment
0
agents auto-scored
Sentiment across 358K calls · one deployment
From signal to action
  • Repeat-caller detection surfaces unresolved issues a one-call metric hides.
  • Contact-rate per merchant exposes upstream problems before they spread.
  • Closed-loop check confirms the order actually moved after the call.
  • High-frustration calls become interventions the agents act on.
Visibility

One control tower. The whole operation.

40s.ai / control-tower
Live snapshot auto-refresh
0
At-risk orders
0
Unassigned
0
SLA breaches
Fulfillment funnel
Live control tower · real-time across thousands of concurrent orders
The 40S briefing

"Right now 847 orders are high-risk, 156 need a driver, and 23 are past SLA. Sharjah is the worst hotspot at 38% off-track. Start there."

AI synthesizes five live data sources into one prioritized instruction, the difference between a dashboard and an operator.

Action

Automate it in plain English.

Describe an operation in a sentence; 40S builds the workflow — agents, conditions, and approval gates — ready to run.

Prompt
Watch it build
The co-pilot

One chat runs the whole operation.

The dashboard is a conversation: ask naturally, 40S answers, renders live artifacts, and routes the specialist work behind the scenes.

40s.ai / dashboard
AAcme Logistics
Last MileFreight
Overview
Dashboard
Workspace
Operations
Predictions
Finance
Reports
Build
Flows
Agents
Data Mapper
Services
History
Afternoon, Sali
Monday, June 29  ·  Keeping it moving.
17.6K orders 116 at risk 1,000 act now 29.7% on time Context

Good afternoon, Sali. I'm 40S, your operations concierge.

19.1K active shipments  |  17.6K orders today  |  97.3% automation rate

Ask naturally. I can track shipments, review risks, explain approvals, prepare reports, and route the specialist work behind the scenes.

What 40S can do

One chat for tracking, live risk, approvals, reports, finance, and customer updates.

Today's operation overview

Get a live snapshot of the entire platform

What can 40S do?

See the main ways to ask for help in one chat

Track an order

Find and trace a specific shipment

Generate a report

Create an operational or performance report

Open canvasPinned artifacts
TrackRiskApprovalsReportFinance
Ask 40S anything about operations… ⌘K↑
Network Focus Canvas Agents
Live network
Active shipments
19.1K
Out for delivery
4.2K
In transit
11.8K
At hubs
3.1K
Hotspots
Sharjah38%
Riyadh22%
Jeddah17%
The live dashboard · operations concierge with a context panel (Network · Focus · Canvas · Agents)
Freight

The same substrate now moves freight.

Last-mile is one hop. Freight is a multi-leg international journey with customs in the middle. We are extending the exact same agent substrate into a full freight forwarding and brokerage suite, across eleven lifecycle clusters, built the way our last-mile suite was.

Origin pickup→Port / airport→Main leg · air, sea, road→Customs clearance→Last-mile handoff
Voice-workflow parity

Run the brokerage by voice

24/7 carrier sales, load coverage, rate negotiation, check-calls and track-and-trace, appointment scheduling, POD and document collection, collections and disputes. Bilingual.

One runtime in place of point tools like HappyRobot, Vooma, Drumkit, Fleetworks.

GCC customs moat

What US tools can't do

Saudi Bayan / FASAH / ZATCA filing, UAE Mirsal 2, HS classification on the GCC unified tariff, landed-cost (duty plus VAT at entry), a demurrage and detention guardian, free-zone and transit.

Region-native forwarding, built on our existing customs pipeline.

Forwarder intelligence

Predictive, not reactive

Transit reality (P50/P90 actual versus the carrier's optimistic quote), schedule search and routing, transshipment-rollover and blank-sailing prediction, equipment and allotment planning.

Powered by licensed market feeds: Xeneta, Sea-Intelligence, project44.

One substrate

Freight reuses the same agent framework, Arabic voice, document-AI, customs pipeline, AP/AR, and FlowForge, retrained on freight and market data. After customs clears, cargo flows straight into the last-mile network. Two verticals, one platform.

Freight · the opportunity

A bigger market, digitizing fast.

Freight dwarfs last-mile, and its digital layer is the fastest-growing part, right as the Gulf builds itself into a global logistics hub.

$270B+
global freight forwarding plus brokerage. The digital layer grows 3 to 4x faster than the market. Grand View, market.us
~27%
CAGR of digital freight brokerage, to roughly $24B by 2030. Grand View Research
10%
of GDP: Saudi Vision 2030 logistics target, backed by $267B+ of investment. vision2030.gov.sa
Why AI, why now: the pain is quantifiable
50-55%
of ocean sailings arrive late (avg 5.3 days). The case for predictive transit. Sea-Intelligence, 2024
630%
ocean rate swing, $1.4K to $10.4K per 40ft. Manual pricing breaks. Drewry WCI
$2,008
per container per day in demurrage, ~$8.9B billed, much of it avoidable. Container xChange
Regional tailwind

Jebel Ali moved a record 15.5M TEU in 2024; Saudi ports handled 320.8M tonnes, up 14.45%. That freight is exactly the lifecycle our agents run. DP World, Mawani

Freight · why we win

The wave is real. It is not built for the Gulf.

Freight-AI is the bright spot of logistics-tech funding, but the leaders are US-shaped and English or Spanish. The Gulf needs Arabic-first, customs-native, and full-stack.

The landscape
PlayerStrengthGap for the Gulf
HappyRobotVoice leader, $62M raised, 70+ brokersEnglish / Spanish, no Arabic, no customs, no last-mile
Vooma · Pallet · FleetworksUS brokerage agents, well fundedUS spot-FTL shaped, no GCC, no customs
Uber Freight30+ agents, global scaleNot GCC-focused, bolted onto a legacy stack
RaftFreight customs automationGeneric global customs, no FASAH / Mirsal
ShipsyRegional, bilingual agent fleetEnterprise-TMS, India-rooted, unfunded since 2022
Our wedge
  • Arabic, Khaleeji-dialect voice welded to freight workflows, on live Gulf operational data. No US leader has it.
  • GCC-customs-native agents that operate FASAH, SABER, ZATCA, and Mirsal 2. No startup does this; Raft is generic.
  • Full-stack: forwarding, brokerage, last-mile, and COD in one suite, the Gulf reality US spot-brokers never touch.
  • Agent-native on real data, not agents bolted onto a legacy TMS.

The honest read: the one to beat regionally is Shipsy, which already markets bilingual agents and customs. We win on operational-data depth, native Khaleeji voice, and owning the full chain through to last-mile and COD, on a substrate already live in the region. a16z, FreightWaves, PitchBook, Shipsy

Act III · Why it wins

Hard to beat,
by design.

Real production proof, a compounding data moat, and intelligence you can try yourself.

Proof · real production data

Not a demo. A live network.

These are real numbers from a live logistics operation running on 40S today.

Delivery success across the network
0
orders delivered or in flight, all-time
Delivered Cancelled / failed
A decade of compounding volume

Annual orders, 2017–2025. The data moat that trains every model.

Where deliveries fail — and what the agents act on
The moat

What cannot be copied.

  • A data moat that compounds. The platform is already trained on 35M+ real orders across a decade, and it sharpens with every order any carrier runs through it. A head start no new entrant can buy.
  • Agent network. 28 agents on a guarded write system. Most competitors have zero agentic automation.
  • Voice, natively bilingual. SIP-native English and Arabic, proven on hundreds of thousands of real customer calls. Most platforms are English-only and text-first.
  • AI-native on live data. Built on a production feed syncing every 30 seconds, not retrofitted onto a legacy stack.
  • Carrier-agnostic. One runtime over any OMS, TMS or carrier. Multi-carrier from day one, not locked to one network.
The compounding moat
Operational data the platform learns from live now a decade of history 35M+ orders already in the platform Every new order, from any carrier, sharpens all 14 models in real time. The moat widens as you scale.
Proprietary intelligence

14 models in production.

Classifiers, by AUC
1.0 0.5 Dispatch timing0.999 Failure prediction0.999 Driver matching0.998 Return prediction0.994 Merchant churn0.986
40s.ai / dashboard
Proactive Risk
1,284 pre37 fix
AWB
Status
Risk
Tier
40SDEMO001
OUT FOR DEL
92%
HIGH
100482214
IN TRANSIT
78%
HIGH
100482007
DROPOFF FA
64%
MEDIUM
100481990
IN TRANSIT
41%
LOW
100481873
PICKED UP
33%
LOW
Live dashboard · ML failure risk, scored per active order
ETA v2 · R² 0.71 Zone demand · R² 0.55 Demand forecast · 73 cities Cashflow + remittance Time-based split, no leakage. The failure predictor powers the autonomous save.
Try it

Score an order's risk in real time.

Adjust the order and watch the failure risk and the recommended autonomous action update live — the same logic 40S runs on every order.

Predicted failure risk
0%
—
Try it

Forecast demand for any city.

Pick a city and see the next 30 days of predicted order volume, with confidence bands. Real model output, 73 cities live.

City
—
avg orders / day
—
peak day
30-day forecast
Competition

Everyone owns a slice.
40S composes the operation.

Generic → Logistics-native Dashboard / EN → Voice + Arabic, agentic Vapi / ElevenLabs FourKites / p44 Locus (IKEA) DispatchTrack Aramex Pallet Raft Uber Freight Vooma Fleetworks HappyRobot Shipsy 40S

Locus routes. FarEye tracks. HappyRobot and the freight-AI pack make calls. Vapi is plumbing. 40S does all of it, one agent runtime on your live orders, in Arabic.

  • Shipsy is the closest in-region, bilingual and customs-aware, but enterprise-TMS, India-rooted, and unfunded since 2022.
  • Freight-AI leaders (HappyRobot, Vooma, Pallet, Raft) are US-shaped and English-first, with no Arabic and no GCC customs. We also out-onboard them: live in days, not quarters, at a fraction of the cost.
  • Global visibility giants (FourKites, project44, Uber Freight) are B2B freight in North America and Europe, not last-mile or Gulf-native.
  • Voice startups are generic plumbing with no logistics data. 40S consumes them.
Act IV · The business

The opportunity,
and the ask.

A large market, outcome-based pricing, onboarding in days, and a clear use of funds.

Market

A growing region, an
exploding category.

GCC last-mile

$14.6B8.6% CAGR$26B

2024 to 2031. Saudi is about 45% of it. Business Market Insights

MEA agentic AI

$102M48.6% CAGR$1.07B

2024 to 2030. Grand View Research

Agentic SCM software

$2Bto 2030$53B

Global spend forecast. Gartner, 2026

Tailwind

Saudi Vision 2030 commits roughly $133B to logistics, targeting the sector at 10% of GDP. By 2030 Gartner expects 50% of SCM solutions to use agents that autonomously execute decisions.

Business model

We charge for outcomes,
not seats.

The structure
  • Subscription floor per tenant, covering isolated fixed infrastructure
  • Usage on top, priced per successful resolution in a $1 to $3 band
  • Voice billed at one all-in rate per minute, no stacked invoice
  • Messaging as pass-through, escalations to a human never billed
  • Bring-your-own-key, so volatile token cost can sit on the customer's P&L
Why usage-pegged, per resolution
price band $1 to $3 Open model$0.10 Gemini Flash$0.15 Gemini 3.5$0.65 Claude Sonnet$1.20

Cost per resolution sits well under the price band. Pegging price to the resolution protects margin when model prices move.

Onboarding

Go live in days, not quarters.

40S sits on top of your existing systems. No rip-and-replace. Four steps from contract to autonomous operations.

1

Connect

Read-only into your OMS, TMS, carrier and comms systems. Oracle, SAP, Salesforce, WhatsApp, any stack.

2

Ground

We ingest your SLAs and SOPs and stand up your isolated data replica and warehouse.

3

Activate

Pick the agent skills you need. They start analyzing immediately, read-only.

4

Supervise

Turn actions on behind approval gates. You watch it work, then widen autonomy.

Connects to your stack
Oracle EBSSAPSalesforceAny TMSWhatsAppEmail / SMSVoice / SIPWebhooksREST / CSV
Per-tenant isolation — your data stays yours
Bring your own LLM keys
Self-serve onboarding, guided by an agent
Go-to-market

Prove it in production,
then expand.

Phase 1

Alpha on real traffic

A live carrier's operators run 40S daily on real orders. Every agent and voice call is instrumented. This is the unfair distribution advantage: a production design partner from day one.

Phase 2

Friendly pilots

A small number of regional 3PLs onboard through the existing customer-onboarding agent. The 20-minute demo becomes the sales artifact carried into every meeting.

Phase 3

GCC expansion

Win the category as the autonomous, voice-first logistics platform, starting in the GCC as the beachhead, then expand by geography and carrier.

Year-1 target: roughly 15 paying customers, 1 to 2 million monthly voice minutes, about 10 million agent calls per month, $1.5 to $2M ARR.

Roadmap

The path to Level 5.

now

Supervised autonomy

Closing the loop on the failure model: autonomous reassignment, dispatch tuning, and proactive customer contact under guardrails.

Q3

Connectors and reach

CRM and tooling integrations, more voice languages for the regional driver base, deeper write-back to carrier systems.

Q4+

Network optimization

Cross-route, cross-driver, cross-hub optimization and external signals: weather, traffic, congestion.

vision

Autonomous operations

40S runs routine operations end to end. The human role shifts from operator to strategist.

Make failed deliveries nearly impossible, not by warning you, but by having already prevented them.

Team

A lean build, by design.

Zaid Chahien

CAIO

Platform engineering, AI agent design, the ML pipeline, the voice stack, and DevOps. Effectively everything below the product line.

Rami Al Hendi

CTO

Strategic direction, product, and the carrier executive interface. Opens the doors and sets the bar.

From two to a launch team
  • Three MVP hires: platform and DevOps, voice and realtime, customer success
  • Carrier-provided support for infrastructure, alpha users, and the commercial wrap
  • Year-1 organisation of roughly 18 across engineering, product, success, and commercial

The platform that exists today, 28 agents, 14 models, a live voice stack, was carried by a compact team using AI as a force multiplier. The raise closes the gap between the scope of the platform and the size of the team.

The ask

Back the operating
system for logistics.

A production platform, a decade of proprietary operational data, and a category opening at exactly the right time.

L4
supervised autonomy, live today
35M+
orders of proprietary data, compounding
$1.5 to 2M
ARR target inside Year 1

Use of funds: the three MVP hires, the voice and reliability hardening, and the first customer pilots.

40S 40s.ai