Skip to content
Back to Blog
Case study Logistics Colombia

How ANT Mobility Is Replacing Phone Calls With a Freight Marketplace and Booking Automation (Logistics, Colombia)

Inside ANT Mobility's freight marketplace in Colombia: SICE-TAC pricing, weighted carrier matching, WhatsApp load offers and 11 systems now in final acceptance testing.

October 21, 2026 · 9 min read
Share:
How ANT Mobility Is Replacing Phone Calls With a Freight Marketplace and Booking Automation (Logistics, Colombia)

ANT Mobility is a Colombian freight platform that connects shippers, carriers and individuals who need to move cargo. Since May 2026, WebMinds, an AI and software development company based in Atlanta, Georgia, serving the US and Latin America, has been its engineering partner. Together we turned a phone-and-chat operation into a freight marketplace with booking automation: eleven systems, from a SICE-TAC pricing engine and weighted carrier matching to WhatsApp load offers, an AI agent and e-signed trip summaries, delivered as one coordinated platform that is now in final acceptance testing before launch.

Key facts

  • 11 systems in one platform: pricing, marketplace, Express, onboarding, WhatsApp, trips and more.
  • 4 portals: shippers and carriers, B2C Express, ANT operations and a driver link.
  • 53 of 53 tables under row-level security, with 88 policies and server-side authorization on every action.
  • 8 AI features in the product, each with its own model, guardrails and usage metering.
  • About four and a half months from first commit (May 22, 2026) to final acceptance testing.

Context: an asset-light freight broker running on calls and chats

ANT is an asset-light freight broker. It owns no trucks; it connects companies that move cargo with third-party carriers on corridors like Bogotá – Medellín – Cali – the Coast. Before the platform, the operation ran on a web form, chat messages and people. Quoting, assigning and following up on a load was manual, prices were negotiated case by case, and nothing was traceable.

The brief asked for a WhatsApp-first core with AI that understands natural language and voice notes, automatic document validation through OCR, and pricing based on SICE-TAC, the official freight cost reference table published by Colombia's Ministry of Transport.

The challenges we started from:

  • Manual operation: quotes by form and chat, assignment by phone and zero traceability.
  • Regulatory compliance: prices anchored to SICE-TAC and validated company, vehicle and driver documents (RUT, Chamber of Commerce registration, SOAT and licenses).
  • Three audiences with different needs: shippers, carriers with fleets, and individuals who want to move something big without negotiating.
  • Carriers live on WhatsApp, not on a website, so the platform had to reach them there.

Why a generic load board or TMS wasn't enough

Load boards and transport management systems exist, and many work well in their home markets. ANT's model needed something they don't combine out of the box.

Prices had to be anchored to a Colombian regulatory table rather than set freely. Every company, vehicle and driver had to be validated against local documents before operating. Individuals needed a consumer-style Express flow with online payments through local methods, while shippers needed offers, counter-offers and bulk loads. And the main channel for carriers is not a web dashboard; it is WhatsApp, including voice notes.

Bolting a WhatsApp bot onto a generic marketplace would have split identity, pricing and audit trails across systems. We chose to build one platform where the marketplace, the pricing engine, the WhatsApp channel and the back office share the same data and the same rules.

What we built: a freight marketplace with booking automation

Since May 2026 we have owned product, architecture, development, QA and deployment end to end, working through four rounds of acceptance testing with the ANT team. The work spans our web and app development and integrations and platforms services. These are the core systems.

SICE-TAC pricing engine

The engine applies official corridor rates and intra-city tariffs, together with ANT's own price bands and alerts. An AI advisor suggests bands from market data, while the calculation itself stays deterministic.

Load marketplace with weighted carrier matching

Shippers publish loads and carriers respond with offers and counter-offers, and an award is atomic. An AI "smart load" feature turns a free-text description into a structured load, alongside bulk CSV upload and recurring loads. A weighted matching engine ranks carriers and explains its choices with AI.

The award itself is a single database transaction: the winning offer, the losing offers, the assigned load and the new trip are written together, with conditional updates so two concurrent requests can never both win.

B2C Express with AI quoting and online payments

For individuals, AI reads the request and the deterministic engine sets the price and signs it. The price travels as an HMAC-signed token, so the browser cannot alter price, vehicle or weight. Payment runs through Bold or Wompi (PSE, card and Nequi), and the paid shipment then becomes a load in the marketplace.

Self-service onboarding with OCR

A 5-step wizard collects company, vehicle and driver documents. Google Vision OCR reads them, Gemini extracts the data and cross-checks it against what was typed, and an AI pre-qualification helps the ANT team, without ever replacing the human approval.

WhatsApp-first operations and an AI agent

Carriers receive targeted load offers on WhatsApp with Accept, Reject and Counter buttons. An AI agent understands text and voice notes and hands off to a human when it should. Identity comes from the Meta-verified phone number, never from what the model says. Voice notes are transcribed with Groq Whisper, Gemini works through 11 tools filtered by that identity, and write actions need explicit confirmation and are audited. It is a key pattern for AI assistants that act on real data: the agent can only do what the user is allowed to do.

Trips without an app

Trip summaries are e-signed and versioned by fingerprint. Each PDF is generated on the server with a SHA-256 hash; if the vehicle or the driver changes, the fingerprint changes and new signatures are required. Drivers use a magic link to sign and report milestones with photo and location, with live tracking and incidents, and no app to install.

ANT back office

The operations team has 39 screens for approvals, loads, trips, tariffs, matching weights, WhatsApp monitoring, per-feature AI settings, usage observability and a full audit trail.

Under the hood

ANT is not a logistics website. It is a system of business rules, atomic transactions, integrations and scheduled processes. The interface is the visible half of a backend of roughly the same size: about 48,000 lines, including 8,000 of SQL, 69 versioned migrations, 32 database triggers, 117 audited action types and 9 roles across four portals.

Every external door is idempotent. Payment webhooks, WhatsApp messages and load offers are claimed with unique indexes before they run, so retries from Meta, Bold or Wompi never execute twice. The platform has four signed webhooks (WhatsApp, Resend, Bold and Wompi) and 28 transactional email templates.

We build with Hybrid AI Coding: senior WebMinds engineering owns architecture, judgment and every decision with the client, while AI agents (Claude Code) handle implementation with independent QA on every change. That is how eleven systems reached final acceptance testing in about four and a half months.

Running freight, field services or any operation coordinated by phone? Get a free assessment and we will map the first systems worth automating.

Timeline

  1. May 2026: proposal, 19 clickable mockups and 28 use cases; MVP approved.
  2. May 22, 2026: first commit. Within ten days: OCR onboarding, AI pre-qualification, the SICE-TAC engine, matching and signed trip summaries.
  3. July 2026: full back office, Bold and Wompi payments, B2C Express end to end, WhatsApp offers and the AI agent, plus a full security audit.
  4. August 2026: first acceptance-testing round with the client.
  5. September 2026: ANT Mobility brand, SaaS-style Express portal and public landing; rounds 3 and 4 close the driver link and versioned signatures.
  6. October 2026: final acceptance testing and launch preparation.

Impact so far

ANT Mobility is pre-launch, so the impact today is in what the platform makes possible rather than in operating volumes.

From phone calls to one platform

Quoting, assignment, signatures, tracking and payments live in one system with a full audit trail, replacing forms, chats and manual follow-up.

Compliance by design

Prices are anchored to the official SICE-TAC table, every company, vehicle and driver document is validated before anyone operates, and data-protection retention under Colombia's Law 1581 is built in.

Meeting carriers where they already are

Load offers, actions and driver work happen on WhatsApp, with an AI agent that understands text and voice and hands off to a human when needed. Screens and details are in the ANT Mobility case study.

FAQ

What is SICE-TAC and why does it matter for a freight marketplace in Colombia?

SICE-TAC is the official reference system for freight transport costs published by Colombia's Ministry of Transport. ANT's pricing engine anchors corridor and intra-city prices to it, so quotes start from the regulatory reference instead of case-by-case negotiation.

Can carriers really work entirely from WhatsApp?

That was the design goal. Carriers receive load offers with Accept, Reject and Counter buttons, can talk to an AI agent by text or voice note, and drivers sign trip summaries and report milestones through a magic link, with no app to install.

How do you stop users from tampering with an AI-generated price?

The AI only interprets the request. The price comes from a deterministic engine and travels as an HMAC-signed token, so the browser cannot change price, vehicle or weight before payment.

Is ANT Mobility live?

Not yet. As of October 2026 the platform is in final acceptance testing and launch preparation, after four rounds of acceptance testing with the ANT team.

Have you built other logistics software?

Yes. For international freight we built an AI document engine for Bills of Lading and Packing Lists in the Kila DocGen case, and for agricultural supply chains an EUDR-ready traceability core in the Loski case.

Build the platform your operation runs on

ANT Mobility shows how far a phone-based operation can go when pricing, matching, WhatsApp and compliance share one backend. Explore the full ANT Mobility case study for the screens and architecture.

Get your free assessment or book a call with our team.

Ready to explore these ideas?

Let us help you turn insights into action. Get a free, no-obligation assessment of your digital strategy.

Book a Free Assessment→
Stay in the loop

Want more insights?

Subscribe to get weekly insights on strategy, technology, and growth delivered straight to your inbox.

Hi! I'm Atlas