unthinkable
Logistics & mobility tech · AI Product Engineering

Help fleets get every load and every rider there on time, without anyone wondering where they are.

For companies building logistics and mobility software

We help you build the product that makes it possible: AI that watches every trip against the plan, tells shippers and riders first when it changes and reads the paperwork so nothing waits on it, with rates and approvals running as code.

The freight work a product can take on

39%of truck stops end in detention: dock time a free-time clock can surfaceATRI, 2024
$725Mlost to US cargo theft in 2025, up 60%: the risk carrier-identity checks are there to reduceVerisk CargoNet, 2026
135 millionhours drivers were detained at docks in 2023, time a product can record as it happensATRI, 2024
What we build

What we build for logistics and mobility companies.

Three kinds of logistics and mobility product we build again and again. For each: how it gets loads and riders there on time, what AI does (and what people still decide), and how we help you build it.

Fleet & logistics (transport, dispatch, analytics)

Every load delivered on time, and shippers and dispatchers hear about a delay before they have to ask.

BUILD AI IN

Agents watch trips against the plan, flag delays, speeding and route changes, read PODs and rate confirmations, and draft the customer update and recovery options. Dispatchers decide the fix.

BUILD WITH AI

Connected to telematics devices, carrier APIs, EDI and your TMS; rates and tolerances run as code. You own every line.

Shipped forDrewry, Autoload, Youngs

Public transit & mobility (live tracking, ticketing, ride-sharing)

Riders know when the bus or ride will come and pay without a queue; operators see every vehicle live.

BUILD AI IN

Live arrival times from every vehicle, digital tickets, alerts when a vehicle or device goes quiet, and rider questions answered in the app. Operators decide schedules and fares.

BUILD WITH AI

One layer that standardises data from every GPS device and ticket machine, built to stay up at peak hours, on cloud you control.

Shipped forChalo

Supply chain & assets (inventory, assets, procurement)

The right stock in the right place on time, and everyone knows what has been ordered without chasing.

BUILD AI IN

Stock and expiry alerts, supplier quotes compared against policy, purchase orders prepared for approval, and supplier email turned into updated records. People approve every order.

BUILD WITH AI

Built on your ERP and WMS data, with scanning that works offline and syncs later; approval limits run as code.

Shipped forSpares Superhero, Suuchi

Running a logistics business or a fleet? Logistics & fleets →

Your customers' journey

The same shipment journey, as your product sees it.

Six steps between a shipper's request and a delivered load. At each: what your product does, what it has to connect to, and what a fleet or broker buyer will check before signing.

Today: a wait before every stepWith AI at work: straight through

1 · Quote

Your productParses the quote request from email or portal, looks the lane up in the rate agreement and returns the price; off-contract lanes go to a rep's queue.

Connects toTMS rate tables and the shipper's contract, plus the inbox or portal the request arrives in.

Fleet and broker buyers checkThat contract rates come from the agreement by code, and the model never makes up a price.

Fleet & logistics (transport, dispatch, analytics) ↓

2 · Booking

Your productCompares carrier authority, insurance and contact details with official and vetting-service records, then matches driver, truck and tracking device to the contracted carrier.

Connects toOfficial carrier records, the operator's vetting services and the load record in the TMS.

Fleet and broker buyers checkThat a mismatch blocks the booking for a manager to decide, and the product never accuses a carrier.

Fleet & logistics (transport, dispatch, analytics) ↓

3 · Pickup

Your productStarts a free-time clock at geofence arrival using each contract's terms, and alerts the shipper before detention begins.

Connects toGeofences, ELD or app tracking, check-call replies and contract terms held in the TMS.

Fleet and broker buyers checkThat detention clocks start from recorded times, not estimates.

Fleet & logistics (transport, dispatch, analytics) ↓

4 · Tracking

Your productAnswers status questions from the latest position and texts the driver only when data is stale or a stop runs long, reading free-text replies, voice notes and photos.

Connects toELD and mobile-app tracking providers, SMS, and the load in the TMS.

Fleet and broker buyers checkThat an arrival time is never shown without fresh tracking data behind it.

Fleet & logistics (transport, dispatch, analytics) ↓

5 · A delay

Your productRecomputes arrival from live position, the driver's legal hours and the appointment, then messages shipper and consignee with the reason and the new time.

Connects toTracking feeds, hours-of-service data from the ELD, and dock appointment scheduling.

Fleet and broker buyers checkThat estimates respect each driver's legal hours, and the product never asks a driver to drive faster.

Fleet & logistics (transport, dispatch, analytics) ↓

6 · Delivery

Your productExtracts signatures, dates and exception notes from POD and BOL images, matches them to the load and routes shortage or damage notes to a person.

Connects toCarrier apps, email inboxes, scanners and the TMS document store.

Fleet and broker buyers checkThat each field carries a confidence score, and low-confidence fields go to a person for review.

Fleet & logistics (transport, dispatch, analytics) ↓

What the product does at each moment (text version)

Quote

  • Read the request (Email, portal or text) → Price from the contract (Lane rates and fuel table in the TMS) → Off-contract lanes (Priced by a rep, never by the model) → Build the load (TMS order with PO and dock notes) → Carrier sales covers it (Carrier and buy rate). Guardrail: Lanes outside the contract, or with unusual requirements, go to a rep to price before the shipper sees a number.

Booking

  • Dispatcher books the carrier (Rate confirmation) → Identity checks (Authority, insurance, contacts) → Mismatch stops the booking (Before any rate confirmation) → Tell the shipper in plain words (Who is coming, and when) → Share release details (Driver and truck to the dock). Guardrail: A carrier that fails a check is never booked; a manager decides what happens next.

Pickup

  • Truck arrives (Geofence or driver check-in) → Free-time clock (Contract terms by facility) → Warn the shipper in plain words (Reason for the wait, options) → Shipper's dock decides (Door order and loading) → Record release time (For the POD and any charge). Guardrail: The agent never promises to waive detention; that is a person's decision.

Tracking

  • Understand the question (Load number, PO or 'my Atlanta load') → Read tracking data (ELD, driver app or carrier updates) → Estimate arrival (Distance, hours of service, appointment) → Customer service for exceptions (Late, off route, temperature alarm). Guardrail: If tracking is stale, the agent says so and asks the carrier rather than estimating.

A delay

  • Exception reported (Driver, tracking or carrier) → Impact on the appointment (Arrival vs window) → Tell the shipper early (Plain words, options) → Customer service rebooks (Receiver call). Guardrail: Appointment changes are confirmed by a person with the receiver.

Delivery

  • Delivery (Geofence and driver upload) → Read the POD (Signature, counts, notes) → Match and file (Load record, temperature limits) → Customer service on exceptions (Shortage, damage, refusal) → Release to billing (Only when the POD is clean). Guardrail: A POD with a shortage, damage or refusal note goes to a person; the agent never reports it as a clean delivery.

Try it with your own case →

Why products stall

Why logistics products stall, and how we help you avoid it.

BUILD AI IN
“The AI will set a wrong rate or approve the wrong spend.”
Rates, tolerances, approval limits and compliance checks run as code. The model reads and drafts; it does not set prices or approve spend, and anything above a threshold goes to a person.
“It can't read our documents.”
Scanned PODs and emailed rate confirmations vary widely. We measure extraction accuracy per document type on your own documents and route low-confidence fields to a person.
“It will cost too much per shipment.”
We batch document work, use smaller models where accuracy allows and report cost per shipment or per document during the pilot, so you can price the feature.
BUILD WITH AI
“It won't fit our TMS, WMS and ERP.”
We connect TMS, WMS, your ERP, carrier APIs, EDI and email, with retries and reconciliation so nothing is lost between systems.
“Our devices and partners all send data differently.”
We build one layer that turns every GPS device, carrier feed and file drop into one standard format, monitored so you see at once when a source stops sending.
“We can't hire a logistics-tech team fast enough.”
A senior lead and a complete team from week one, with building blocks already proven in production: connectors, an exception queue, an action log and reconciliation reports.
You inherit

Building blocks already built.

Already built
Document extraction for BOL, POD, invoices and customs formsin production
EDI and carrier API connectorsin production
Rate and tolerance rules enginein production
Exception queue with approval by rolein production
Email-to-record agentin production
Action log and reconciliation reportsin production
Your logistics & mobility tech productstarts on these, not from scratch

The team around it

Product designerShapes how people and agents work together
AI engineerAgents, prompts, evaluations
ArchitectRecords, integrations, security
DevelopersBuild it with AI, keep it maintainable
TestersSpecs and tests that prove it works
Domain expertKnows your industry's rules and edge cases
Security reviews

Will your product pass a fleet's or broker's review?

Before any pilot, operators ask how a product handles rates, carrier identity, drivers and their data. These are the answers we build in from the first release.

BUILD AI IN
Prices and spend stay in code
Rates, tolerances and approval limits run as code; the model reads and drafts, and anything above a threshold waits for a person.
Identity checked before money moves
Carrier authority, insurance and contacts are compared with official records at booking, and a change of bank details always needs a person.
A log an auditor can replay
Every agent action is logged with its inputs, the rule that applied and who approved it, so an error can be traced and reversed.
BUILD WITH AI
Driver hours respected
Arrival estimates use each driver's legal hours from the ELD; the product never asks a driver to drive faster or skip a break.
Data kept to its load
Shipper, driver and contact data is used only for the load it belongs to, and never to train models.
Sandbox and test loads first
Built against the TMS sandbox and test loads before live freight; each confirmed mistake becomes a test case for the next release.
Integrations

The integrations fleets and brokers ask for.

An operator will ask what your product connects to before it asks about features. These are the connections we build.

TMS, WMS and ERP

TMSs such as McLeod, Turvo and Aljex, plus WMS and ERP data, through their APIs, with records reconciled between systems.

Carriers and partners

Carrier APIs, and EDI where partners require it; email and file drops where nothing else exists.

Tracking and devices

ELD and mobile-app tracking, telematics and GPS devices brought into one standard format, monitored so a silent source shows at once.

Vetting and paperwork

The carrier-vetting services operators already use, and PODs, bills of lading, rate confirmations and customs forms read and matched to loads.

You own the result: code, prompts, evaluation sets and data sit in your accounts from the first line, so your product's IP stays yours.

Proof

Logistics and mobility products we have shipped, and what they did.

Chalo, India's bus-tracking and ticketing app40%less time spent managing data, across 15,000+ public vehicles in 17 cities

Android e-ticketing devices and one backend that syncs routes, stops, duties and ticket revenue, with live vehicle tracking for operators and passengers, on AWS. Ticketing and GPS hardware from many vendors didn't work together, so dispatch teams lacked live trip details and fares leaked.

Read the case study →
Chalo, India99.9%uptime, under 2 seconds average ping from each device

Middleware that turns every device's data into one standard format, with an operator dashboard showing which devices are live, delayed or silent. 15,000+ GPS devices sent data in different formats and at different intervals, and a delay of even 5 seconds made it useless.

Read the case study →
Asian Paints, India85%fewer in-transit damages (as reported, from a pilot on four routes)

IoT devices in vehicles that flag potholes, rough roads and aggressive driving in real time, with a road-condition map from plant to customer and trip reports. Paint was damaged in transit by road conditions and driving, and road surveys were manual and ran only every three months.

Read the case study →
A UAE asset-tracking providerresult to confirm

A modernised platform: a live map of vehicles, alerts for speeding, route deviations and maintenance, driver hours and performance, and reports on fuel and utilisation. Its fleet platform could not track shipments in real time, monitor speed or send instant alerts.

Read the case study →
Suuchi, a US apparel supply-chain company28%lower manufacturing overhead, 46% faster order processing

One cloud platform from sample to shipment: costing, raw materials, procurement, quality, inventory, production and invoicing, with a panel for customers to track status. Fragmented communication across customers, factories and mills delayed product development and production.

Read the case study →

Results as reported in our client case studies.

Which delay would your product remove first? Name the step where loads, orders or riders wait most: booking, dispatch, tracking or proof of delivery. After a short discovery you get a plan with dates for the first release, measured against that delay.

Tap one: Ask Unthinkable starts your plan from there. Or say it in your own words ↓

How to choose

How to choose who builds logistics & mobility tech products.

Four things a logistics-tech founder or CTO should check, and how we meet them.

Has shipped logistics products

Fleet, transit and supply-chain products for companies such as Chalo, Autoload and Suuchi, and a fleet platform for a UAE asset-tracking provider.

Often elsewhereStrong engineers who are new to telematics, EDI and freight paperwork.

Knows the operator's side

We also help brokers, fleets and 3PLs run day to day, so we know the dispatch, dock and billing work your product has to fit.

Often elsewhereLogistics-only shops know the domain; few have built AI that acts.

Agents inside the product

In a fleet product, for example: trips watched against the plan, PODs and rate confirmations read, and the customer update drafted for a dispatcher to send.

Often elsewhereA chat widget bolted onto a tracking map.

Feature twenty as fast as feature one

We set the architecture before the prototype and keep AI writing inside it; rates and tolerances stay in code, so a rule change is a tested release.

Often elsewhereA fast demo that stalls once EDI, devices and carrier feeds arrive.

The full comparison: us, other software companies, in-house teams and consultants →

Ways to work

Start the way that suits you.

Time & materials

A senior AI team billed for time, within an agreed monthly budget. Even on time and materials the work is tied to your business objective: every plan starts from it, and every report shows the number next to the hours.

One fixed outcome

One agreed outcome at a fixed price, with acceptance criteria, including the measure, written up front.

Dedicated team

A complete senior AI team, monthly, working as part of yours and reviewed each quarter on what shipped and what moved.

Linked to the result

Where the gain can be measured from your own systems: a reduced implementation fee plus a percentage of the expected gain, for a set period.

How we work and price →

Straight answers

Questions logistics-tech teams ask.

How do I choose a development partner for a logistics or fleet-tech product?

Test them on the parts that break logistics products. Ask how they would bring mixed GPS devices, carrier feeds and file drops into one format, and how they notice when a source goes silent. Ask how they connect a TMS, WMS, ERP and EDI partners without losing records between them. Ask how they measure document-reading accuracy on your own PODs and rate confirmations, and what happens to a low-confidence field. Ask whether rates and approval limits will stay in code, and whether they report running cost per shipment so you can price the feature. Finally, check that code and data sit in your own accounts. Unthinkable has shipped Chalo's GPS middleware, with 99.9% uptime across 15,000+ devices, and its e-ticketing platform, and it also helps fleets and 3PLs run, so it knows what your customers will ask.

How can AI be used in logistics software?

The most useful place is exceptions and paperwork. Agents read freight documents, match them to loads, spot late shipments, audit invoices and draft customer and carrier messages. Rates, tolerances and approvals stay in code, and planners approve rate changes and exceptions. Tracking itself is usually already solved; the time goes into the work around it, and that is where agents help.

Can you integrate with our TMS, WMS and ERP?

Yes. We integrate through APIs where they exist, EDI where partners require it, and email or file drops where nothing else is available. We build and test in a sandbox first, reconcile records between systems and log every write, so your operations team can trace what changed.

How accurate is AI at reading freight documents like PODs and bills of lading?

Accuracy varies with scan quality and layout, so we measure it on your own documents before launch rather than quoting a general number. Each field carries a confidence score; low-confidence fields go to a person for review. Over time, corrections become new test cases. Most teams see the review queue shrink as the mix of document types becomes known.

How much does it cost to add AI agents to a supply-chain product?

Integrations usually drive the cost more than the AI. Most projects start with one workflow, such as freight audit or document processing, as one fixed outcome, with a written estimate after a short discovery. Unthinkable measures running cost per document during that first release and reports it per shipment, so you can price the feature.

I run logistics operations and have a product idea. Can you build it?

Yes. You know the process and where time is lost; we bring the complete product team: product designer, AI engineer, architect, developers, testers and a senior lead. We start with a short discovery to scope a first release and fix the price. You own the code, IP and data, and we stay until it works with real loads.

What happens when an AI agent makes a mistake in a shipment workflow?

Mistakes are contained by design. Agents can only take actions within limits set in code; anything above a threshold, such as a rate change or a large order, needs a person. Every action is logged with its inputs, so errors can be traced and reversed. Each confirmed mistake becomes a test case, so the same error is caught before the next release.

Can you modernise our legacy logistics platform without stopping operations?

Yes, by replacing it in parts rather than all at once. Unthinkable first captures how the current system behaves as tests, then moves one module at a time, such as booking, tracking or billing, while old and new run side by side and records are reconciled between them. Each module switches over only once it passes the same tests, so loads keep moving throughout.

Ask Unthinkable

Where do loads or riders run late in your product today?

Tell us about your product and the moment you want to change. Our AI assistant, Ask Unthinkable, helps you define your requirements, then connects you with a senior AI engineer.

Send us a note instead

Required

Thanks. A senior AI engineer will reply to you.
u
Ask UnthinkableClarity first, then the right team
Your conversation

An AI agent. This prototype uses prepared answers and sends nothing.