The physical-truth layer for irregular freight

Freight is priced
on a guess.
We measure it.

Every quote, load plan and compliance call in road freight rests on dimensions someone estimated in a yard. Stereon turns messy physical items into verified, reusable, quoteable, loadable freight assets — and feeds every system that already exists.

See the thesis PILOT LIVE · VISION INTELLIGENCE
SCROLL
01

Information integration

Design files, 3D scans or rough dims — every item becomes one verified freight asset, DG declared at the door.

02

Truck optimisation

Verified envelopes packed onto the right vehicle — fewer trucks, fuller decks, no auto-approvals near a limit.

03

Carrier-side logistics

Dual-regime pricing, DG-gated documents and a signed load plan the carrier can actually trust.

01 — The problem

Freight moves on physical data nobody actually has.

Carriers bill the greater of dead weight and cubic weight — and at Australia's 250–333 kg/m³ conversion, a hand's width of dimensional error moves the invoice ~24 kg. On irregular freight, every number in that equation is a yard guess. Everyone in the chain pays for it differently.

Worked example — real design-partner item
ItemCamera kit · wrapped pallet
Measured dims1.2 × 0.8 × 1.1 m
Volume1.056 m³
Actual (dead) weight60 kg
Cubic weight @ 250 kg/m³0 kg
Carrier bills0 kg
Shippers

Pay for air and book trucks they don't need

Guessed dims → cubic penalties, 30–40% quote-to-invoice variance, and a second vehicle "to be safe." Messy loads travel unspecified — then get damaged or rejected.

Carriers

Booked blind

Freight lands bigger than declared: depot re-measures, wasted deck space, missed permit lead-times, DG surprises at the dock — and the driver wears the compliance risk.

3PL / 4PL

Forced to price on someone else's guess

They quote off what the customer tells them — every re-bill erodes margin, every dispute erodes the relationship, and awkward freight gets priced defensively or turned away.

End customers

"Will it fit, and what will it cost?"

Movers and retail deliveries answer that question when the truck is already in the driveway. Wrong vehicle, second trip, surprise charges — all for want of dimensions known up front.

02 — The gap

Freight software knows everything about the shipment — except the shipment.

The commercial data layers are mature. The physical ones barely exist. That bottom band is the entire wedge.

Customer & ordersERP / TMS · good
Carrier ratesTMS / broker · good
Tracking & invoicescarrier APIs · adequate
Dimensions / cubic profileyard guess · poor
Load fit / geometry"will it fit?" · very poor
Restraint / DG / oversizePDF rules · inconsistent

Every existing load-planning tool assumes you already know the item's dimensions and will type them in — which is exactly the step that fails for irregular freight. Stereon attacks the bottleneck itself: capturing physical truth at source.

03 — The product

Three ways in. One verified truth. Every truck optimised.

An item enters however it exists — an engineering drawing, a phone scan, or a rough measurement — and becomes a versioned library asset with packaging, handling and DG attributes. The library optimises the truck, then hands the carrier everything they need — live in the pilot today.

1

Scan & identify the item

2

Optimise the truck

3

Verify the regulations

4

Download the load plan

SCANNING…
LIVE GEOMETRY ENGINE

Scan & identify the item

Point a phone at it — the scan resolves the object and its verified dims land in the library as a reusable freight asset. No scanner on hand? Dims come straight off the engineering drawing (the AI reads it, an operator confirms) or as rough dims, flagged until re-measured. Packaging and orientation set the true shipped envelope — a panel travelling vertical on an A-frame eats a pallet space, and the system knows it.

Optimise the truck

The engine rearranges the load for pallet-space and deck-rate optimisation — collision, stacking and orientation rules, deck metres, with axle loads checked after every placement and every drop: a load that is legal at departure can overload an axle at drop three. The objective is trucks removed, not utilisation percentages — and every arrangement is one a forklift driver can actually execute.

Verify the regulations

The loaded truck is swept against versioned, per-state rule packs: mass and dimension limits, restraint feasibility per the Load Restraint Guide, axle groups, oversize thresholds. WA runs different rules to the HVNL states — interstate lanes bind to the most restrictive jurisdiction on the route. Near any limit it escalates to a human. It never auto-approves.

DANGEROUS GOODS — DELIBERATELY SEPARATE

Declared at input, before an item is placed: no UN number + class → no DG documents. Segregation per ADG Table 9.1, placard thresholds computed per leg, protective placement enforced around batteries. Stereon screens and documents; it never pretends to certify.

Download the load plan

One click: a signed plan per vehicle — top-down placement diagram, axle profile at every drop, restraint checklist, pricing regime and surcharge flags, stamped with the rule-pack version it was checked against. That audit trail is what a Chain-of-Responsibility duty actually needs: evidence, not certification.

04 — The moat

The scanner is not the moat.
The library is.

Photogrammetry can be wrapped by anyone. What compounds is the customer-specific library of verified freight assets — geometry, confirmed weights, handling flags, version history, and the audit trail of every quote and plan built on them.

L6

Switching cost

Leaving means re-digitising the entire catalogue and retraining operations.

L5

Benchmark data

Cross-customer geometry and utilisation benchmarks no one else can assemble.

L4

Compliance trail

Rules plus audit logs quietly become the freight desk's risk shield.

L3

Integrations

Plans and dimensions flow into the TMS layer — Machship, Cario and beyond.

L2

Data

Item libraries accumulate verified physical truth with every scan.

L1

Workflow

Warehouse teams touch it daily — repeat items are instant, new items are a two-minute scan.

05 — The market

A precise wedge into a very large cost pool.

Buyers already pay US$15K–100K per fixed dimensioning unit where the ROI is obvious — proof the budget exists. Stereon is mobile, works on oversized freight, and produces a reusable asset instead of a one-off number.

US$0B Australian road freight — the immediate cost pool Stereon optimises ~67% of AU domestic freight task
US$0B Global TMS software by 2030 — the layer Stereon embeds into from US$18.5B in 2025 · ~15% CAGR
US$0B Automated dimensioning by 2033 — the budget line that already exists fixed hardware today · mobile tomorrow
Now

Beachhead shipper

Vision Intelligence — freight so awkward it breaks the manual status quo weekly. Prove savings on the hardest case.

Next

Repeat-irregular verticals

Traffic management, fencing, solar, equipment hire, staging & AV, modular buildings — same pattern, same library economics.

Then

TMS embedding

One Machship integration reaches 100+ 4PL operators. Stereon becomes the physical-truth module of the platforms freight already runs on.

Scale

International 4PL

DB Schenker, DHL, Geodis — the identical irregular-freight problem, worldwide. The library model travels.

06 — Who wins

One physical-truth layer. Four ways to extract value.

The same verified geometry means something different — and something valuable — to every party that touches the freight.

Owner-drivers & small fleets

Maximise every load

Know exactly what fits before saying yes. Verified envelopes turn "she'll be right" into a packed deck — and make taking on extra freight a calculation, not a gamble.

Shippers

Less trucking, tamed loads

Messy, irregular consignments become specified, safely handled, correctly restrained freight. Fewer "second truck to be safe" bookings, fewer damage claims, fewer re-bill disputes.

3PL / 4PL

Lease the layer

White-label physical truth into their own offering: sharper quotes on freight competitors won't touch, instant re-quotes from the library, and an audit trail their customers can lean on.

Retail & moving

The consumer scanner

A customer scans the couch, the fridge, the flat-pack. The mover or carrier knows the item, its dimensions and what the job actually needs — before the truck is booked.

The backload play: a meaningful share of truck-kilometres run empty on return legs. Today "can you fit it on the way back?" is answered by phone tag and guesswork. A shared library of verified envelopes makes it an instant yes/no with a price — that's where this layer goes once the library is dense enough.

07 — The pilot

We're proving it on the ugliest freight we could find.

Vision Intelligence — design partner

Assembled camera towers, concrete counterweights, gate sections, pole bundles, lithium-battery kits — dispatched interstate weekly with no TMS and manually booked consignments. A freight profile almost perfectly adversarial to the status quo, which is exactly why it's the proving ground.

● PILOT PLATFORM LIVE ● SHARED ITEM LIBRARY DEPLOYED ◐ CAPTURE VALIDATION UNDERWAY
50–80irregular units dispatched weekly
7interstate lanes, MEL to PER
Class 9lithium batteries — UN3480 DG flow
2pricing regimes: cubic and heavy-haul
We run this like an experiment, not a pitch. Capture accuracy on the hardest items is a hypothesis under live validation against tape-measured ground truth, with pre-committed kill / continue / accelerate gates. If the physics doesn't hold, we'll say so first. That discipline is the point — it's what makes the data investable when it lands.
08 — Research & diligence

We pressure-tested the thesis before writing a line of code.

A feasibility assessment, not a pitch deck. Five questions had to survive honest scrutiny — each with a stated confidence, and the caveats left in.

QuestionFindingConfidence
Is the problem real & expensive?
Chargeable freight is priced on dimensional data that, for irregular items, is guessed — driving mispricing, disputes and over-trucking.
High
Is the gap genuinely open?
3D load planning exists, but nobody starts from irregular-freight capture, builds a reusable geometry asset, prices it against live AU rates, and wraps it in compliance.
High
Is the technology ready?
Ready enough to test, not yet proven for freight. Must be validated on the hardest real items before it's assumed — which is exactly what the pilot does.
Medium-high
Is the market big enough?
Australia alone is a modest vertical-SaaS niche. Venture scale rests on TMS embedding and international expansion. Sized honestly, three ways.
Medium
Is there a credible exit?
Strategic acquisition — TMS platforms, WiseTech, international 4PLs — at sober multiples, conditional on proven metrics.
Medium-high

Pricing mechanics

Cubic-weight conventions (250 / 333 / 167 kg/m³), the deck-metre and heavy-haul regimes, and the full accessorial cost stack — mapped to where geometry actually moves the invoice.

Competitive white-space

Load-planning SaaS, fixed dimensioning hardware, phone dimensioners and TMS platforms — each mapped to the precise capability none of them own.

Measurement tolerance

Use-case tolerance bands, confidence-driven behaviour, and a Phase-0 capture protocol with pre-committed kill / continue / accelerate gates.

Market sizing

TAM / SAM / SOM built bottom-up, cross-checked against global TMS software and the automated-dimensioning hardware market.

Regulatory boundaries

NHVR Chain of Responsibility, the Load Restraint Guide, the ADG Code for lithium batteries, and IATA air rules — scoping what the product may and may not claim.

Risk register

Eleven risks scored on severity × likelihood, each with a mitigation — capture accuracy, false precision, adoption, liability creep and TAM honestly among them.

Grounded in public industry data — Gartner supply-chain surveys · Australian carrier pricing conventions · NHVR & ADG regulation · academic 3D bin-packing literature · private-SaaS M&A comparables. Full viability assessment available under NDA.

09 — The vision

Physical truth is the last unowned data layer in freight.
We intend to own it.

Validation pilot in market now. Evidence pack — accuracy tables, variance analysis, operator adoption — lands at the end of the 90-day sprint.

Contact us
10 — Team

The team.

KM

Kavan Mehta

Founder
LinkedIn ↗
JL

Jayden Lay

Co-founder
LinkedIn ↗