# Case study: a pricing engine the client's own finance team could trust

How Equipoise turned a tangle of spreadsheets into a fast, validated pricing tool, and proved every number before it went live.

## The situation

A property-services firm priced complex back-office service deals by hand, across a sprawl of spreadsheets that only a few people fully understood. Quotes were slow, hard to reproduce, and easy to get wrong. They needed a tool that anyone on the team could use to price a deal consistently, and that leadership could actually trust.

## What Equipoise did

- Reverse-engineered the firm's existing pricing logic into a clear, documented specification.
- Designed and directed the AI-enabled build of a pricing and financial-modeling application: portfolio inputs, staffing and cost models, margin and scenario analysis, and clean client-facing and internal outputs.
- Iterated with the firm's stakeholders over many rounds as the model and the business questions sharpened.

## The part most people skip: verification

Rather than ship a tool that merely looked right, Equipoise built a second, independent model whose only job was to check the first one's math, cell by cell; reconciled both against the firm's own internal spreadsheets, line by line, until the numbers matched to the dollar; and ran structured quality passes that surfaced and fixed real defects before they could reach a customer quote.

## The outcome

The firm got a fast, self-serve way to price deals consistently, with the math verified against ground truth rather than taken on faith. The engagement also produced a clear path to a production web app for wider rollout. Equipoise stayed through the iterations and left the firm owning the tool, not dependent on a black box.

## The takeaway

Anyone can generate something that looks like a solution. The work is proving it is correct, scoping it to your business, and leaving it in your hands. That is the difference between AI-enabled delivery and AI slop.