Case studies

Real systems, explained in plain words

Every project below is in production. Client names, prices and contract terms are left out on purpose. Everything else is here: the problem, what we built, how it works and what changed.

Freight · Quotations

The quotation bot: a quote written from one email

For a Canada-based international freight forwarder.

The situation

Customers ask for ocean freight prices all the time, and many expect a refreshed quote every two weeks. Each quote meant opening that customer's workbook, finding the current rates across several source files, copying them over, checking them and formatting the result.

It was careful, skilled, repetitive work, and it competed with everything else on the desk.

What we built

A bot that takes its instruction from an email. A team member forwards the request. The bot works out which account it is for, opens that account's quotation workbook and the source rate files, and produces the quote in the layout the customer already knows.

What changed

The team reviews a finished quote instead of building one from scratch. Quotes follow the same steps every time, whoever asked for them, and the two-week cycle no longer piles up.

In plain words

You send an email saying what you need. A finished price quote comes back for you to check.

Freight · Rate management

Carrier rate sheets into one rate master, automatically

For the same forwarder's pricing team.

The situation

A forwarder buys space from shipping lines and resells it. Each shipping line sends its prices as a spreadsheet, each in a different layout, and each updates often. The pricing team kept one internal master file, and bringing every new sheet into it was done by hand, line by line.

One wrong row means a wrong price in front of a customer.

What we built

A pipeline that reads the carrier's sheet, including refrigerated cargo rates, and turns it into clean data. It works out the service for each line from the carrier's free-text routing notes, then matches every line to the right row in the master using several fields together, so similar-looking lanes are not confused.

Fixed rules do the matching. AI is used only for the unclear parts, and it is shown column headings, never the prices.

What changed

A new carrier sheet becomes an updated master in one run. The team's time goes into checking the exceptions the pipeline flags, not retyping the rows it got right.

In plain words

Shipping lines send price lists in messy formats. The system reads them and updates the company's own price list, then shows a person anything it is not sure about.

Freight · Contracts

A carrier contract turned into a working rate workbook

For the same forwarder's pricing team.

The situation

A contract with a shipping line is a long document full of rates, surcharges and conditions. Before anyone can quote from it, it has to become a spreadsheet the team can use. When the contract is amended, someone has to find what changed.

What we built

A conversion that turns the contract into a structured rate workbook in a fixed layout. A second step compares the new workbook with the previous one and lists every difference: new lanes, removed lanes, changed rates.

Along the way it repairs the small spreadsheet faults that break comparisons, such as stray empty columns.

What changed

Pricing starts from a clean workbook and a short list of changes, instead of reading two documents side by side.

In plain words

A long contract goes in. A tidy spreadsheet of prices comes out, with a note of everything that is different from last time.

Freight · Tracking

Container tracking across nine carriers

For the same forwarder's customer service team. Rolling out carrier by carrier.

The situation

To tell a customer where their cargo is, staff logged in to each shipping line's website and searched container by container. With nine carriers, that is nine websites, nine logins and nine different screens.

What we built

One tracking design for all nine. Where a shipping line offers a direct data connection, we connect to it. Where it does not, the request goes through an aggregator, a service that already collects tracking data from many lines. Either way the answer comes back in the same shape.

Where it stands

The first carrier is live. The second is being connected. The rest are queued, some waiting on the carrier's own approval process. We say this plainly because that approval, not the software, sets the pace.

In plain words

Instead of checking nine websites to find a container, staff look at one screen.

Healthcare · Auralyn

Unifying a multi-center diagnostics network

For a medical clearance and diagnostics operator in South Asia.

The situation

Registrations were disconnected, clearances lived in paper books, and no one had a single source of truth across centers.

What we built

A diagnostic-center operating system covering patient intake, test orchestration, clinician sign-off and certificate generation. One queue per role. One record per patient, across every location. Every change tied to a user and a time.

What changed

Paper clearance books were retired. Turnaround on medical clearance certificates went from days to hours, with a verifiable trail for every record.

In plain words

A patient is registered once. Every test, payment and signature after that is recorded in one system, and the certificate prints at the end.

See Auralyn
Finance · Pledgium

A branch-scale lending platform with live ledger control

For a gold-backed lending network in South Asia.

The situation

Manual ledgers in every branch, no real-time view at head office, and a painful reconciliation every week.

What we built

A multi-branch lending platform with configurable loan products, valuation checked by two operators, and a live cashflow ledger. Branch managers see positions in real time, head office sees every branch rolled up, and auditors see the full life of every ticket.

What changed

Reconciliation moved from a weekly exercise to continuous. Exceptions dropped once every manual entry had to be confirmed by a second person in the system.

In plain words

Every loan against gold, in every branch, is recorded the same way, and the cash in the drawer is checked against the system all day.

See Pledgium
Freight terms

The jargon, translated

Freight has its own language. Here is what the words in these case studies mean.

Freight forwarder
A company that arranges shipping for businesses. It books space with shipping lines and handles the paperwork.
Carrier
The shipping line that owns the vessels and moves the containers.
Rate sheet
A carrier's price list: what it costs to move a container from one port to another.
Rate master
The forwarder's own combined price list, built from every carrier's rates.
Lane
One route, such as a specific origin port to a specific destination port.
Reefer
A refrigerated container for cargo that must stay cold. Priced separately.
Quotation
The price a forwarder offers its customer for a shipment, valid for a set period.
Demurrage and detention
Fees charged when a container stays at the port, or with the customer, longer than agreed.
Confidentiality

What we never publish

Our clients trust us with their pricing. These pages are written so that anyone can understand the work without learning anything a competitor could use.

  • ×

    Client and carrier names

    Unless a client has asked to be named.

  • ×

    Rates, margins and contract terms

    No real price appears anywhere on this site.

  • ×

    Internal files and templates

    We describe what a file does, not what is in it.

Want the longer version?

Under NDA, we can walk through the architecture and the problems we hit on the way to production.