Case studySystems that don’t talk

Quotes in seconds, not an hour

On a sales team, how long a quote takes often comes down to how experienced the person is — because most of the work isn't quoting, it's hunting for the numbers.

Before → now30–60 min / quoteInstant
How it fits togetherFive places to check. One quote.Price, spec and availability came from separate systems. Now they flow into the quote on their own.
Sector
Sales & distribution
The before
Hunting across siloed databases for every line item
What we built
A tool that pulls the data into the quote instantly
Stack
FastAPI · PostgreSQL · HTML · Local LLM

At a glance

  • Building a quote dropped from 30–60 minutes to near-instant.
  • Ask a built-in assistant in plain language — no combing the database.
  • Speed no longer depends on who happens to be doing it.

The situation

The sales team worked across many databases that didn't talk to each other. Building a quotation meant a long fact-finding exercise — chasing the price, the spec and the availability for each line item across separate systems. A quote took anywhere from 30 minutes to an hour, and how fast depended entirely on how experienced the salesperson was.

What we did

We built a tool that connects those siloed databases and brings the data into the quote instantly. Instead of hunting system to system for every line, the salesperson gets what they need in one place — so the quote comes together in seconds, whoever is building it.

It also has a built-in assistant: rather than combing the database, the salesperson just asks — in plain language — for a price, a spec or the right line, and pulls the answer straight into the quote. Quick in, quote out.

What changed

A quote that took 30 minutes to an hour is now near-instant. The fact-finding is gone, the numbers are consistent because they come straight from the source systems rather than memory, and a new salesperson can quote as quickly as the veteran.

Why it worked

The slow part was never the quoting — it was the data living in separate systems. Once they're connected, the quote builds itself.

Quotes taking an hour of fact-finding?

We connect the data behind the quote, so it builds in seconds.

Show us the problem