Skip to content
samuel_pedro
// we complement2025
$ cat ~/case-studies/we-complement.md

Quotes that took hours now take under two minutes

A paraplanning firm was pricing every job by hand across five service domains. I mapped the pricing logic and built a calculator that applies it consistently and exports a client-ready quote.

// client
We Complement
// role
Fractional AI Lead
// year
2025
// stack
Next.js

Context

We Complement provides paraplanning to UK financial advice firms. Fees are fixed rather than hourly, which sounds simpler than it is: the price depends on which of five domains a case touches, how complex it is, how many products are involved, and which discounts apply when domains combine.

The problem

All of that logic lived in people's heads and in spreadsheets. Quoting a job meant reconstructing the rules from memory each time. It was slow, it varied between advisors, and an arithmetic slip landed in front of a client as a price.

The firm did not need better spreadsheets. It needed the pricing rules to exist in one place, outside anyone's head.

What I decided

The first real work was not building anything. It was sitting with the pricing logic until every rule was written down explicitly, including the ones nobody had articulated before, like exactly when a multi-domain discount applies and what happens when it collides with a product surcharge. Most of the value was created in that mapping. The build was the easy half.

I built it in Bubble rather than writing an application from scratch. For an internal tool with a handful of users and pricing rules that were certain to change, the ability to alter logic in an afternoon mattered more than anything I would have gained from a custom stack.

The interface is three steps, in a fixed order: configure the rules, select the case, generate the quote. I resisted making it flexible. A pricing tool that lets you skip ahead is a pricing tool that produces inconsistent quotes, which is the exact problem it exists to solve.

What I did not do was extend the tool backwards into intake. Cases still arrive by email and get entered by hand, and the calculator picks up only once someone is ready to quote. That was the right call for a first build, because solving pricing was the problem the firm actually felt, but it means the tool sits inside a manual process rather than replacing one. The intake automation is a separate piece of work, and the calculator made the case for it.

What shipped

A pricing calculator with the full fee structure encoded, automatic multi-domain discounts and product surcharges, a visual breakdown showing where the cost sits, and PDF export producing a branded quote. Desktop and tablet. Tested with advisors before rollout.

The result

Quotes are generated in under two minutes and the arithmetic errors are gone, because there is no arithmetic being done by a person. The less obvious gain is consistency: two advisors quoting the same case now reach the same number, which is what makes a fixed-fee model defensible to a client.

// result

What it produced.

< 2 min
to produce a quotewas hours
5
service domains encodedpricing logic, written down
0
arithmetic errorsno person does the sums
2025
deliveredfractional ai lead
// related

This started as a fractional ai/technical lead.

You need someone owning the technical calls week after week - architecture, hiring bar, what gets bought instead of built.