Tocnis designs, builds, integrates and operates custom autonomous agents for freight forwarders. Every agent is engineered from the workflow up, lives inside your existing stack, and runs under live observability — so your team gets capacity back and your executives can see exactly what the AI is doing.
Every freight forwarder we speak with describes some combination of these three problems. They sound like operational issues. They are actually symptoms of the same gap: nobody is automating this industry with the depth it deserves.
Extracting data from bills of lading. Replying to status questions. Building quotes from carrier sheets. Reconciling invoices line by line. Filing customs documents. The work is structured, repeatable — and burning through your FTE budget.
You've tried a SaaS tool that never quite worked. An OCR product that broke on the 30% of fields it couldn't read. An agent platform that demanded months of internal configuration. Generic tools assume generic workflows; freight is anything but.
Leadership is being asked to hand bills of lading worth six figures, invoices that hit the GL, and customer communications to a system they can't audit. That hesitation is rational — and the single largest barrier to AI adoption in this industry.
Tocnis builds autonomous agents that execute specific high-frequency workflows end to end — with human oversight where it matters and full autonomy everywhere it should not.
Engineered from your document formats, your systems, your SOPs and your edge cases — not from a generic template. Prompts, evals and reasoning chains belong to you.
Agents read PDFs, emails, EDI feeds and portals — and write through your TMS, ERP and CRM with scoped credentials. No new app for your team to learn.
Delivery is the start of the relationship. We monitor, refine against new edge cases, upgrade models and report against agreed metrics — under one named human.
Each agent has a role, a model, and a place in your reporting structure. An orchestrator routes work; specialists execute; every task carries a chain of custody you can audit.
Every freight operation has its own bottlenecks, but most cluster into a small number of recurring patterns.
Bills of lading, commercial invoices, packing lists, customs forms — read, extracted, validated, written to your TMS. No keystrokes.
Carrier sheets, contract rates, surcharge tables. LCL vs FCL, ocean vs air. Customer-ready quote in seconds.
Inbound quote requests parsed from email and portals, missing detail gathered, and a priced response drafted — ready to send or auto-release.
Carriers contacted for capacity and pricing, quotes requested and rates negotiated against your targets — escalated only when terms fall outside your mandate.
Status pulled from carrier portals, EDI feeds and email. Exceptions, delays and milestones detected and routed.
Booking confirmations, check-calls and status follow-ups run with carriers automatically — every reply logged back to the shipment file.
Declarations prepared from shipment data. HS code mismatches and valuation issues flagged before they hit a broker.
Carrier invoices compared against booked rates. Discrepancies flagged line by line. Approvals routed to the right person.
Status questions, booking confirmations, ETAs. The agent knows when to answer and when to hand off to a human.
The connective tissue. Coordinates handoffs across teams — sales to ops, ops to customs, ops to finance — and enforces the SOP. Sits on top of every other agent.
Built to remove ambiguity, surface risk early, and give you measurable value at every checkpoint. How long each stage takes depends on the workflow and the agent.
Map the workflow, the systems it touches, the edge cases and the metric that will measure success.
Architecture, decision logic, integration points and observability surface — in writing, with success criteria.
Built against real production data in a controlled environment. Iterated until it beats spec on the real edge cases — not the synthetic ones.
Monitor, refine, upgrade and report. The agent gets better every month.
OngoingMap the workflow, the systems it touches, the edge cases and the metric that will measure success.
Architecture, decision logic, integration points and observability surface — in writing, with success criteria.
Built against real production data in a controlled environment. Iterated until it beats spec on the real edge cases — not the synthetic ones.
Monitor, refine, upgrade and report. The agent gets better every month.
A fixed project fee for building each agent, plus a monthly retainer for operating it. Two levels of ongoing involvement, and a low-risk way to start small.
| Start small Starter engagement. One agent, one workflow. Prove the model in your operation. | Option B Hybrid handover. We build, your team operates. Tocnis handles the deep changes. | Most popular Fully managed. Tocnis owns the agent end to end. Your team uses it. We run it. | |
|---|---|---|---|
| Agent engineering | |||
| Number of agents | One | Multi-agent | Multi-agent |
| Named account lead | |||
| Monthly retainer | |||
| 24/7 monitoring | |||
| Model upgrades | Your team | ||
| Major changes | Scoped | ||
| SLA | Defined | Defined | Defined + |
| Team training | |||
| Best for | A low-risk pilot before expanding. | Operators with technical staff who want the agent in-house. | Operators who want AI value without standing up a team. |
One agent, one workflow. Prove the model in your operation.
A low-risk pilot before expanding.
We build, your team operates. Tocnis handles the deep changes.
Operators with technical staff who want the agent in-house.
Tocnis owns the agent end to end. Your team uses it. We run it.
Operators who want AI value without standing up a team.
Fixed build fee + monthly retainer. Full proposals are shared after Discovery — you see exactly what you're paying for and what we're committing to. Pricing reflects scope, not optimism.
Tocnis is led by a founding team supported by senior technical advisory expertise. Every engagement gets direct access to specialists — not layers of generalists.
Nicolas is a recognized specialist in digital twins and AI agent engineering, with a long track record of building, deploying and operating autonomous systems in enterprise environments.
His role at Tocnis is not symbolic. He's involved in the technical architecture of every client engagement, the review of every major build, and the recruitment of every engineer who works on a Tocnis project — giving us the bench depth of a much larger firm.
The founding team owns strategy, client engagements and commercial direction, with deliberate focus on building one of the most disciplined AI engineering practices in the sector.
Senior and intermediate engineers are recruited through our Technical Advisor's network, vetted for freight-specific delivery experience, and engaged against the scope of each client project — bandwidth of a much larger firm, without the overhead.
Eight of the most common questions we hear in a first call — answered plainly, the way we'd answer them on the phone.
It depends on the agent and the workflow — the complexity of the process, the number of edge cases, and the depth of system integration all shape the timeline, so we don't quote a fixed number that fits every project. What's consistent is the path: Discovery, Design, Build & Integrate, then Operate — with a clear deliverable and measurable value at each stage before the next begins.
Yes. Our entire model is built around meeting your systems where they live. We work with the major freight TMS platforms as well as custom and homegrown systems. Integration scope is confirmed during Discovery.
Every engagement starts with agreed success metrics, written into the Discovery deliverable. We measure and report against those metrics throughout the engagement. Typical metrics include hours recovered per week, throughput per FTE, error rate, exception rate and cycle time.
Every Tocnis agent runs against a defined boundary of authority. Inside that boundary, it acts. Outside it, it escalates to a human. Every action is logged. When something goes wrong, we can trace exactly what happened, why, and what to change.
We treat agent errors the way a serious engineering team treats software incidents: structured post-mortems and concrete fixes.
Yes, and we recommend it. Our Starter engagement is built exactly for this: one agent, one workflow, one measurable outcome, on a defined timeline. It's the lowest-risk way to see how we work and judge the value for yourself.
If you run operations at a freight forwarder, logistics provider, customs broker or 3PL, we'd like to talk. The first conversation is short and unfiltered — we want to understand the work that's slowing your team down, the systems you already run on, and the outcomes that would matter most. From there we'll tell you, honestly, whether Tocnis is the right partner.