If you run freight on CargoWise, Magaya, SAP, Oracle, GM Transport, or Excel, hear this: you are probably paying for a tool that only works when your team feeds it by hand.

TMS tools are glorified databases at the core: someone captures the shipment, uploads evidence, and moves status. If the team does not, the system knows nothing. That is the human bridge between systems: licenses bought, work still manual.

OCL Cargo is different at the root: it is an autonomous TMS. A TMS that works alone with agents and computer use that complete workflows end to end.

This pillar explains computer use in logistics, how it differs from RPA/API, and when a pilot makes sense at 300+ shipments/month.

If your TMS only works when someone feeds it

In real corridor operations, a large share of visible ops time goes to capture, manual tracking, and repetitive tasks that do not add margin.

That is not a lab study: it is WhatsApp, the carrier portal, and the TMS. It is not the tool's fault; it was designed to register, not to execute.

If truth lives in Excel or a TMS nobody updates on time, name the pattern. Then choose another screen or agents that operate the ones you have.

Buyer hub: best TMS Mexico 2026 (alternative to TMS tools that only register).

Register vs execute: the difference that matters

System of record

  • Stores orders, rates, and status
  • Depends on manual capture
  • Alerts only if someone builds them
  • Sample audit or parallel Excel

System that executes

  • Operates portals and screens (computer use)
  • Assigns, tracks, and evidences without asking
  • Useful alerts in minutes (GPS mirror)
  • Audits 100% CFDI + Carta Porte + GPS + POD
OCL does not require replacing your record at the outset: it adds execution on top.

Register asks: what was captured? Execute asks: who did the work at 2 a.m. when the tractor stopped before Laredo? You search for a TMS; the pain is execution.

What computer use is (and is not)

Computer use is the agent's ability to use the computer with an analyst's goal, without fatigue or forgetfulness.

  • Navigate a TMS, GPS portal, inbox, or WhatsApp Web
  • Read the screen and choose the next click
  • Validate the result and leave a trail in the shipment file

It does not invent rates. It works with your rules (rate card, alert SLA, tax checklist) and evidence. Citable definition: computer use in logistics.

Computer use vs RPA vs API / iPaaS

API / iPaaS

How it works: Moves data between systems with connectors

Typical Mexico freight failure: Carrier has no API; TMS does, GPS does not

When it is enough: Homogeneous stack and integration budget

Classic RPA

How it works: Fixed UI macros

Typical Mexico freight failure: One portal button change breaks it

When it is enough: Ultra-stable processes, few exceptions

Computer use

How it works: Agent operates UI with trip context

Typical Mexico freight failure: Bad rules or incomplete evidence

When it is enough: Multi-portal, WhatsApp, mirror accounts, no API

On Mexico–US networks, carriers rarely expose the API your RFP imagined.

Mexico–US context: Laredo, CFDI, Carta Porte

A Monterrey–Laredo shipment does not close on a green status. You need complete evidence:

  • Freight CFDI
  • Coherent Carta Porte 3.1 complement
  • GPS evidence (ideally with a geofence at dock or bridge)
  • Usable POD (photo, signature, accessorials)

If that lives in four inboxes, you pay blind.

OCL agents verify CFDI and Carta Porte in buyer-side audit. They are not a PAC: they do not issue or stamp.

Checklist: CFDI + Carta Porte + GPS + POD audit.

What OCL agents execute end to end

  1. Assignment

    Voice and WhatsApp to carriers; system record, not chat-only.

  2. Tracking

    GPS mirror accounts; useful alert in under 10 minutes.

  3. POD evidence

    Dock, app, or WhatsApp tied to the trip.

  4. Audit

    Match CFDI, Carta Porte, GPS, and POD.

Four agents, one file.

If you are deciding to complement or replace a system of record: decision playbook and 6-8 week playbook.

When it fits (and when it does not)

Elige un paso para ver el detalle

Detalle del paso · 01

300+ shipments/month

Signal 1

Manual capture no longer scales without headcount.
If you check 3 or more, the bottleneck is execution, not another TMS license.

OCL also administers own fleet and workshops: the playbook fits shipper, 3PL, and operators with assets. If volume is very low and disciplined capture already covers 100% with no leakage, pilot ROI takes longer to show.

Cluster and related reading

Pillar: this article. Supports and TMS cluster:

Key takeaways5 points
  1. A typical TMS (CargoWise, Magaya, SAP, Oracle, GM Transport, or Excel) records what the team captures; without capture, the system knows nothing.
  2. Computer use lets agents operate software, portals, and screens without waiting on complex APIs.
  3. It is not brittle RPA or iPaaS: the agent runs the shipment workflow.
  4. On Mexico–US freight the file needs CFDI, Carta Porte 3.1, geofence/GPS, and POD matched before payment.
  5. At 300+ shipments/month, a 6-8 week pilot often frees capture time and recovers 5-7% of spend via audit.

Frequently asked questions

If you run 300+ shipments/month on the Mexico–US corridor, see agents on real shipments and measure a 6-8 week pilot.

See agents on real shipments