A mid-sized logistics company in the United States replaced a manual, spreadsheet-driven pricing process with AI-powered tools. Analysis that took days now completes in minutes, and the team moved from quoting one customer at a time to more than twenty concurrently.
The problem
Every customer quote required building a custom rate card from scratch.
An analyst would pull shipment history, work through volumes and service levels by hand, and construct pricing in a spreadsheet. For a single customer this took days. Across a sales pipeline it became the constraint on how fast the company could quote — and therefore on how many customers it could pursue.
The spreadsheet dependency carried a second cost. Rate cards built manually across dozens of tabs, by different people, at different times, produce errors. In pricing, an error is not a cosmetic problem. It is either revenue left on the table or a contract that loses money for its full term.
The company was not short of data. Shipment records, service levels, customer history — all of it existed. What did not exist was a way to turn that data into a priced, defensible rate card without a person spending days on it.
What we built
Automated rate card generation
The system ingests a customer’s shipment profile — volumes, lanes, service levels, dimensional characteristics — and generates a complete rate card. What required a multi-day manual build now runs in minutes.
Shipment analytics
Alongside pricing, the tooling surfaces the analysis that used to be assembled by hand: shipment patterns, cost drivers, service level distribution, and the profile characteristics that determine whether an account is worth pursuing and at what price.
Both tools were built to be operated by the people who need them, not by analysts. The parcel teams use them directly. The build took approximately four weeks from engagement to working tooling.
What changed
From one customer at a time to twenty-plus
This is the change that mattered commercially. The manual process allowed the team to work one customer at a time. The automated tools allow more than twenty concurrently. Opportunities previously deprioritized — because the analysis cost more than the account appeared to be worth — became viable to quote.
A 25% revenue growth opportunity
The increase in quoting capacity opened a potential revenue uplift of roughly 25%, driven by pipeline the team could not previously reach rather than by any change to pricing itself.
Errors designed out
Pricing is generated from source data through a consistent process rather than assembled by hand across spreadsheets. The category of error that comes from manual transcription and inconsistent method is removed rather than caught later.
Adopted across the organization
The tools are in daily use by all parcel teams. That matters more than it might appear — analytics tooling that gets built and then ignored is the common outcome. These are operational.
Why it worked
Two things, neither of them the model itself.
The data foundation came first. Before any automation, the shipment data had to be reconciled, consistently structured, and reliable enough to price against. Most of the early work was there. A pricing tool built on inconsistent data produces confident, wrong rate cards — considerably worse than a slow manual process.
It was built for the operator, not the analyst. The people who needed rate cards were not the people who could build them. Closing that gap is what turned a technical result into an operational one.
Project highlights
- Mid-sized logistics company, United States
- AI-powered rate card generation and shipment analytics
- Fully automated; replaced a manual spreadsheet process
- Analysis time reduced from days to minutes
- Concurrent capacity increased from 1 customer to 20+
- Opened an estimated 25% potential revenue growth opportunity
- Built and delivered in approximately 4 weeks
- In production use across all parcel teams
Is this your problem too? If your pricing, quoting, or customer analysis runs on manual spreadsheet work, the constraint is rarely effort — it is that the process cannot scale past the number of people doing it. Talk to us about automating it.