Case study
We built Spencer.
Spencer is the AI model that helps truckers navigate the legal system built around trucking. Carso built it for T.A.T.T.O.O. Foundation — a 501(c)(3) nonprofit serving truckers. Live in production at tattoo.foundation.
Below the Spencer case study, two more projects: a personal AI knowledge graph that turns your hard drive into something searchable with Obsidian, and SuperTrucker — the fleet AI platform we built for ourselves.
Context
Truckers rarely have a lawyer.
T.A.T.T.O.O. Foundation (Truckers Against Trafficking — Truckers & Owner Operators Foundation) exists to serve truckers navigating the legal system. Labor trafficking. Predatory lease-purchase arrangements. Wage theft. DOT disputes. Compliance actions. Most truckers face these without counsel and without a clear starting point.
Generic legal chatbots don't work for this — they don't know trucking law, they don't speak to truckers like truckers, and they don't understand the operational reality behind a compliance question. The foundation needed something built for the job.
What we built
An AI model trained for one job.
Spencer is an AI model fine-tuned on the legal system that surrounds trucking. Carso designed, trained, deployed, and integrated it into the foundation's infrastructure.
Trained on trucking law.
Spencer reads the FMCSA regulations, the ECFR, state trucking statutes, and the case law that surrounds them. Not a general LLM with a trucking prompt bolted on — an AI model specifically built to know how the legal system treats trucking.
Handles the legal system.
When a trucker faces a labor-trafficking situation, a compliance dispute, a contract question, or a regulatory action, Spencer is a first point of contact. It answers plainly, points to citations, and knows when to escalate to human counsel.
Built for truckers.
Truckers rarely have a lawyer on retainer. Most legal help written for them is written by people who've never been in a cab. Spencer changes that — instant access to legal guidance, in language truckers actually use.
The proof
Spencer knows every regulation.
We fine-tuned the Spencer AI model on the full text of the FMCSA regulations and the Electronic Code of Federal Regulations. No other AI model in trucking can say that.
It means Spencer answers DOT questions with citations. It catches violation patterns generic AI misses. It knows what an officer actually checks during a roadside inspection. Generic ChatGPT guesses at trucking law. Spencer trained on it.
Mission work
Built at no cost to the foundation.
Carso donated the design, engineering, training, and deployment. This is the kind of work we do when a mission aligns with ours — bringing real AI to the industries that need it most.
See Spencer live at tattoo.foundation.
Project — personal AI knowledge graph
Turn your hard drive into a knowledge graph with Obsidian.
Most operators have a decade of files scattered across their hard drive — receipts, contracts, customer threads, project notes, regulatory PDFs, scanned paperwork. None of it searchable. None of it linked. Carso builds you a personal AI knowledge graph that ingests the lot, structures it inside an Obsidian vault, and lets an AI agent answer questions against your own files.
Local AI models. Your data stays on your machine. You own the vault and the data inside it. We hand it to you and walk away — same way we ship every Carso project.
Your hard drive, ingested.
Years of files, notes, contracts, receipts, customer emails, project folders — we ingest the lot. PDFs, Word docs, screenshots, Markdown, plain text. Nothing leaves your machine. Local AI models do the heavy lifting so your private data stays private.
Structured into Obsidian.
We build you an Obsidian vault — the open-source knowledge tool used by researchers, lawyers, and operators who refuse to lock their notes inside someone else's SaaS. Files become linked notes. Notes become a graph. Concepts surface that you forgot you ever wrote down.
Searchable by an AI helper.
An AI agent sits inside the vault. Ask it a question — what did I tell that customer last March, what's in the contract I signed in 2022, what notes do I have on this regulation — and it answers with citations to the source files. Your own files, finally findable.
Platform — SuperTrucker
The fleet AI platform that replaces four bills.
SuperTrucker is the platform we built for the carrier we wish had existed when I drove freight. One unified stack that takes the place of Samsara, DAT, Motive, and JJ Keller — pretrip AI helpers, compliance AI agents, load-board intelligence, ELD audit. Live in beta at supertrucker.ai.
Same stack we use for client work. SuperTrucker is the proof that the AI infrastructure we ship for clients runs in production every day — not as a demo, not as a slideshow, but as a working platform with carriers using it.
Pretrip AI helper.
Voice and vision AI model that walks drivers through a DOT-grade pretrip inspection and reads OEM-level diagnostic data. Catches what a tired driver misses at 4 AM in a yard.
ELD audit AI agent.
Retroactive HOS audit that catches the violations introduced by ELD departments themselves — the leading cause of DOT audit failures most carriers don't realize they have.
Load-board intelligence.
AI agents that watch the load boards, surface lanes that match your authority, your equipment, and your historic margins. Connected to Common Ground for cross-company freight discovery.
Want us to build your AI model?
Custom-trained AI models, AI agents, AI platforms — built for your specific operation. Scoped in one call. Priced to your company size.