All posts
Photo: Ollie Craig — Pexels
Perspective
Denis Konoplev3 min read

Cross-border still runs on four inboxes and a hope

Most cross-border operations still run on four inboxes and a hope.

The assumption is that fragmentation is the price of the lane — three carriers, three portals, three spreadsheets, and somebody senior who knows which thread to pull when a shipment goes quiet. That isn't a temporary state waiting for the right TMS. It's how the work actually gets done, and it scales until it doesn't.

A parcel from a warehouse in India to a doorstep in the UK passes through air freight, customs clearance, and last-mile delivery — usually three companies, three systems, and three data formats that don't talk to each other. Every handoff is where visibility drops, cost creeps in, and something breaks. Forwarders and 3PLs know this better than anyone. Their options have historically been binary: lock into someone else's stack, or spend years and millions building from scratch.

We think there's a third option, and this post is the thesis behind it. You'll get the commercial model, the product family that ships against it, and where we're taking it next — the thesis, the product family, and what's next.

The thesis

RSA XB is a modular, plug-and-play network for cross-border logistics. Our customers launch their own branded end-to-end service across air freight, customs clearance, and last-mile delivery — and they choose which legs to run on our operations and vetted partners, and which to keep with partners of their own.

Underneath that sits a layer of AI tooling that accelerates the commercial work and automates the operational steps that have always made cross-border slow and inconsistent. A proprietary platform orchestrates the handoffs across every leg and pools volume, so a logistics company of any size can access economics that used to be reserved for the giants.

The point isn't to replace these companies. It's to give them the network and the technology to scale faster than they could alone.

Spun out of RSA Global, backed to scale

RSA XB was carved out of RSA Global, a Dubai-headquartered logistics group, and we emerged from stealth with a $1.5M seed round led by 21 Ventures. We've grown 10% week-over-week for the past twelve months, with an initial focus on the India-to-UK, India-to-Europe, and India-to-GCC corridors.

From our commercial headquarters in Dubai and operations in India, we're using that capital to expand the AI platform, scale India capacity, and deepen our corridor networks.

The product family

The AI tooling isn't abstract. It ships as a family of products that plug into the workflows our customers and partners already run:

  • XB ClearanceFlow — AI-native customs processing that connects to broker systems and automates compliance submissions.
  • XB TrackingBridge — unified, cross-carrier tracking that fills the visibility gaps traditional APIs leave behind.
  • XB AddressIQ — address validation and normalization that cuts failed deliveries.
  • XB Shiprate — rate shopping and quoting across fragmented carrier rate cards.
  • XB Freight — a freight forwarding platform for quoting, booking, and shipment visibility across air, ocean, and ground.
  • XB Alfred — an AI email agent that understands logistics-specific language and automates the inbox.

Each one does a real job on its own. Together, they're the tooling layer for a network that any logistics company can build on.

What's next

Cross-border e-commerce is growing fast, and the logistics sector underneath it is going through a major shift — service providers are getting more sophisticated and going after the global market. We're here to give them the services to create those networks and the technology to scale them.

The next agent won't live in your system of record. It will live in the handoffs — the inbox, the attachment, the carrier event that never mapped cleanly — because that's where the work still happens. Anyone selling you end-to-end visibility by bolting AI onto the platform that records answers after the fact is optimising the wrong layer.