Skip to content
Industry: trucking & logistics

AI automation for fleets that file, haul, and get audited.

Logistics runs on paperwork with deadlines: quarterly fuel taxes, bills of lading, rate confirmations, driver files, safety records. Most of it is retyped by hand, and every number can be checked by a state or a shipper. That is exactly the shape of work this studio automates — and trucking is the one industry where you don't have to take that on faith, because the systems are already in production for real carriers.

automated tests on the IFTA filing agent
429automated tests on the IFTA filing agent
backtest against a real state filing
To the pennybacktest against a real state filing
a real carrier files through it
Quarterlya real carrier files through it
Already running

Not a pitch — systems you can inspect.

Trucking is the industry this practice has shipped the most for. Each of these exists today, with the tests to show for it.

IFTA fuel-tax filing, computed and reviewed

Deterministic Python computes every per-jurisdiction number with state-exact rounding; a Claude agent reviews the finished return against the rule base and 21 quarters of filing history; a human approves before anything is filed. Running in production for a Kentucky carrier — and offered as a service for fleets from one truck up.

Read the engineering case study

Freight document extraction

Bills of lading, rate confirmations, PODs, and carrier invoices arrive as scans and leave as manual data entry. Extraction with field-level validation turns them into structured shipment records — low-confidence reads get flagged for review instead of written silently into your system.

BOL extractor on GitHub

Dispatch, intake, and driver-facing web

Job intake classified and routed by your rules, recurring compliance reports assembled from source systems, and the public-facing side handled too — a trucking company's recruiting site, tested to load in under half a second on a phone, ships from this studio.

More production systems
Where the hours go

Paperwork that repeats, with a deadline attached.

The strongest logistics automations are unglamorous: recurring, rule-bound, document-driven work where a mistake is expensive and detectable.

Quarterly IFTA prepared from raw mileage and fuel files, with an AI review note and a human approval gate before filing.

BOL and rate-confirmation data extracted, validated field by field, and matched to loads — no more retyping from scans.

Carrier and vendor invoice reconciliation where exact matching runs in code and only the exceptions reach a person.

Driver application and intake flows that reach a dispatcher structured, complete, and routed.

Compliance and safety reporting assembled on schedule with every figure traceable to a source record.

Back-office assistants that answer from your own SOPs, contracts, and rate history — with citations.

Why here

Built the way audited industries need.

The same three rules on every logistics system this studio ships.

The math is never generated

A model that is right 99% of the time is wrong on a tax return. Anything with a correct answer — tax math, reconciliation, rate calculations — runs in tested deterministic code. The model reads, classifies, and explains.

Audited-industry discipline by default

Every system ships with automated tests and evaluations, because in this industry a wrong number is not a bug, it is an assessment. The IFTA agent's backtest reproduced a real accepted state filing exactly.

One engineer who knows the paperwork

You work directly with the independent engineer who built these systems — someone who knows what a CDTFA surcharge line and a clean BOL look like, not an account manager relaying notes.

Questions

What carriers and brokers ask first.

Straight answers, including the prices.

Can you automate our IFTA filing?

Yes — it is the most proven thing this studio does. The pipeline ingests your mileage and fuel records, computes a state-portal-ready return with exact rounding, runs an AI review over it, and a human approves before anything is filed. It runs quarterly for a real carrier today, and the first quarterly review is free for new carriers.

We're a small fleet. Is this worth it below 20 trucks?

Usually yes, because the paperwork does not scale down with the fleet. A quarterly filing, invoice reconciliation, or document intake takes nearly the same office time at 5 trucks as at 50. Automation is priced fixed-scope, so you know the number before deciding.

Our documents are photos and scans from drivers. Does extraction still work?

That is the normal case, and it is why extraction ships with field-level validation and a review step. A clean read goes straight through; a low-confidence read is flagged for a person instead of being silently written into your records.

What does AI automation cost for a trucking company?

AI automation projects start at $8,000 as a fixed quote, and smaller data and integration work starts at $1,500. The IFTA filing service is priced separately per quarter. Every price on this site is published, and you get the quote in writing before any work starts.

Do you replace our TMS or dispatch software?

No. Automation connects to what you already run — your TMS, ELD exports, spreadsheets, and inboxes — through their existing files and APIs. The expensive manual work usually lives between systems, and that gap is exactly what gets automated.

Do you work with carriers outside California?

Yes. The carrier running the IFTA agent is in Kentucky, and document and reporting automation runs remotely for any US operation. The studio is based in Sacramento, CA, with on-site work available across Northern California.

Send the paperwork you want gone.

Describe the process — the files, the systems, the deadline — and you'll get an honest read on whether it's worth automating, and a fixed quote if it is.