The marketplace that stays for the work
For a UK logistics client we can't name: a two-sided marketplace matching businesses that need warehousing and transport with providers that have capacity, which then runs the resulting relationship end to end. Stock, despatch, transport, invoicing, settlement. Event-sourced from day one, and carrying the deepest production AI stack we've shipped.
The market, and the problem worth solving
Contract logistics is a long-tail market. Most warehousing and haulage capacity in the UK sits with small regional operators, which makes it expensive to find and slow to buy. Sourcing a fulfilment partner routinely takes months of calls, site visits and negotiation, and usually ends in a multi-year commitment that doesn't flex with seasonal volume.
The supply side has the mirror problem. Operators sit on idle racking and empty return legs, and the cost of winning a new contract through traditional sales is brutal for a business running on thin margins.
Marketplaces are the obvious answer, and mostly they stop at the introduction. The interesting decision here was to keep going: once shipper and provider contract, the same platform runs the operation, from goods-in to proof of delivery to the invoice. That is what turns a directory into infrastructure, and it is why the build is as large as it is.
What we built
One platform, two sides, and a workflow that follows the relationship from first search to settled invoice.
Search that starts anonymous
Public geo-radius search across more than twenty facets, from temperature control to certifications to racking dimensions. An AI concierge qualifies what the visitor needs, shows anonymised matches, and turns interest into a registered, scored lead.
Contracting, line by line
Fourteen service rates negotiated individually, each carrying the provider's list price, both sides' proposals and the agreed figure, plus volume-discount tiers. E-signature is wired into the flow, and the contract state machine won't let a deal skip steps.
An operations backbone
A nine-state fulfilment workflow from despatch note to proof of delivery, enforced server-side, with stock moving through allocated and in-transit counters at every step. Both parties see the same process, trimmed to what each is entitled to see.
A transport quoting engine
Quotes priced from multiple rate-card sources with postcode-zone matching, capacity and dimension gating, and honest labelling: a quote is either firm or indicative, and each carries a cost, capacity and directness score.
Settlement built in
Billing cycles per contract, proforma to finalised invoices with the full margin stack itemised, and payment initiated over open banking. Money flows are part of the platform, which is what makes the commission model enforceable.
Trust as the go-live gate
A listing only appears in public search once vetting passes, and the vetting is done by an AI classifier with domain sense. Suggestions go back to the provider; nonsense never reaches a buyer.
The engineering
This is our largest single codebase: around 380,000 lines across a .NET backend, two front ends and a fleet of AI services, built over more than three years. Two decisions shaped everything else.
Event sourcing, with receipts
Every write appends an immutable event whose body is hashed, and the hash is re-verified on every replay. Twenty versioned migration classes let the schema evolve without ever breaking history. Concurrency is a hash reject rather than a lock, which sounds brave and has proven simple to reason about.
Marketplace tenancy is its own discipline
Standard SaaS isolation gives each tenant its own world. A marketplace can't: both parties to a shipment must see the shared record, each trimmed to their side. Quoting goes further, pricing a shipper's request against many providers' private rate cards through a deliberately hidden reader that can't leak any of them.
Search does the matchmaking
Elasticsearch drives geo-radius search with facet filtering and postcode-aware fuzzy matching, with indexes rebuilt under timestamped names and swapped by alias so a rebuild never takes search down.
Self-hosted where it counts
Notifications, e-signature, PDF rendering and secrets all run in-cluster rather than as third-party SaaS, and error tracking is an in-house system that merges front-end and back-end traces into one issue view. Fewer processors handling commercial data, and no per-seat bills that scale with success.
Tested like it handles money
Because it does. More than 1,200 backend tests, 24 production workloads on Kubernetes in Google's London region, and an admin console with 44 screens for the operations team, including event browsers and replay tooling.
AI with a job to do
This platform runs more AI in production than anything else we've built, and none of it is decoration. Logistics is a trust trade with a paperwork problem, which is exactly the ground where current models earn their keep.
A concierge that can't make things up
The public assistant routes each visitor to a specialist agent for warehouse search, transport quotes or onboarding. Six grounding guards sit behind it, and one of them strips any result the model presents without a matching tool call, so it cannot invent a warehouse. Errors never surface as failures; the chat always answers. Leads arrive in the CRM scored for intent.
Models on trial, permanently
Three generations of the assistant have shipped, running across several providers with weighted A/B testing, an LLM-as-judge evaluation rig with named scenarios, and per-model cost accounting. Which model answers a given turn is a measured decision, not a loyalty.
AI as the front door bouncer
The vetting classifier reads each new listing against domain rules, coordinates that exist, plausible dimensions, credible pricing, and its verdict gates whether the listing goes live at all. The marketplace's quality bar is enforced by a model with taste and a checklist.
Agent-ready before the industry asked
A 50-tool MCP server exposes the whole operational API to AI agents, each tool classified read, create, mutate or destructive. When a customer's agent wants to check stock or raise a despatch note, the platform is already speaking its language.
Shape, effort and money
A programme measured in years rather than sprints, structured so each phase paid for its successor.
Indicative investment
£280k to £450k in year one
A first transactable marketplace of this shape typically runs £125,000 to £170,000. With an operations platform and steady iteration, first-year totals usually land between £280,000 and £450,000, and multi-year programmes of this depth scale well beyond that. Excludes VAT and infrastructure, at London rates as of mid-2026. Shape bands, not an invoice.
Where it landed
The platform is live in the UK market with providers and buyers transacting through it, and the operations side, the part that was hardest to build, is the reason relationships stay on-platform once they start.
It is also quietly ahead of its industry on AI. While logistics still mostly means phone calls and PDFs, this platform vets its own supply, qualifies its own leads and exposes its API to customer agents.
Asked about this project
The engagement is under NDA, which is common for platform builds in competitive markets. The screens shown are indicative recreations in our own visual language, and identifying details, including scale figures the client publishes, have been generalised.
Ready to build the thing?
Book a free 30-minute call. We'll dig into your idea, your stack and your timeline, and give you an honest read on what it will take to build and launch. You'll leave with a clearer plan whether or not you hire us.