The Custom Tool

One tool, for the thing you do four hundred times a year.

You already know roughly what you need. We turn the rough idea into one working internal tool, deployed in your stack, and we get paid out of what it saves.

The arithmetic

Run the sum on your own numbers.

Before we build anything we agree what the loop costs you today, so the honest thing is to show you the sum rather than describe it. Every assumption below is a control you can move, including how much of the work a tool would realistically take over.

Hours a year handed back
What those hours are worth

These are your numbers, not ours. On the call we measure the baseline properly out of your own logs, agree it in writing before anything gets built, and measure it again afterwards.

The point is not the software. It is the hour somebody stops spending on it.

In your stack

Your cloud, your accounts, your login. Not a demo tenancy we control and not a tool you rent from us.

A person stays in charge

Anything that leaves the building, costs money or carries a signature waits for somebody to approve it.

Paid out of the saving

We measure the baseline before and the same number after. If it does not move, you owe us nothing.

The loop

Same request, before and after.

Almost every tool we build collapses a chain like this one. The work does not disappear, it stops being passed around, and the waiting in the middle goes first.

Today
Mail arrives Somebody reads it Waits for the right person Copied into the CRM Price looked up Offer written Sent, then chased
With the tool
Mail arrives Draft ready in 40 seconds A person approves, it goes

Seven touches and two waits become one decision.

The shapes

What we get asked for most.

Different companies, more or less the same six loops. If yours is not here it does not mean no, it means we will want to see it before we say yes.

Inbox to CRM

Mail and forms read on arrival, sorted by intent, written into the system with the fields already filled.

Inquiry to offer

A request becomes a priced draft from your current price list and templates, waiting for a signature.

Documents to data

Invoices, delivery notes and scanned post turned into rows, with anything unclear flagged rather than guessed.

The nightly reconcile

Two systems compared every night, differences listed by morning instead of found in June.

Reports nobody enjoys

The recurring spreadsheet somebody rebuilds by hand every Monday, built once and then just correct.

The handover checklist

Multi-step processes with owners and deadlines, chased automatically, escalated when they slip.

How it goes

Two to four weeks, and you see it running in the first one.

Week 0The call

Thirty minutes with both founders. You describe the loop, we tell you whether it is worth automating and roughly what it is worth. Sometimes the answer is that a spreadsheet formula would do it.

Week 1The baseline, in writing

Before anything is built we agree the number we are going to measure and where it comes from. Everything afterwards is judged against that document, so we would rather argue about it now.

Weeks 1 to 3Build, in the open

You see it running at the end of the first week, ugly and incomplete. Every week after that it does more of the loop, and you tell us where it is wrong.

Week 3 to 4Live, with a person in the loop

It goes into daily use with approval steps switched on. We watch it for the first weeks and fix what real traffic finds.

AfterMeasured, then handed over

We take the same measurement again. Then either we keep running it, or your team does, or we teach them to.

What we need from you

  • One person who does the work today and can be interrupted for an hour a week.
  • Access to the systems in scope, and somebody who can grant it without a six-week ticket.
  • Agreement on the baseline number before we start. This is the only part we are strict about.

That is the whole ask. If nobody can be spared for an hour a week, the tool will be built against what we imagine your process is rather than what it actually is, and it will be wrong.

When we say no

Four times we turn this down.

It happens twice a month.A small saving stays a small saving, and the build still costs the same weeks. Below roughly an hour a week we will tell you it is not worth either of our time.
The process changes every time.If no two cases are handled the same way, there is no loop to automate yet. What you have is a decision that needs writing down first, and that is a different job.
Nobody agrees what the process is.When two people describe the same workflow differently, building software freezes one version of it into place. We would rather send you back to settle that than pick a side by accident.
You want it to decide.We build tools that prepare decisions and stop. If the requirement is that nobody looks at the output before it reaches a customer or an account, we are the wrong studio.
Afterwards

Who runs it once it works.

We doWe keep it running: hosting, monitoring, fixes and the improvements that come out of daily use.
Or your team doesEverything is in your accounts already, so this is a handover rather than a migration.
Or we teach themFour sessions on the running system, quoted separately, and only ever after it has proven itself.
Either way you own itCode, prompts, deploy scripts and documentation, in your repository, from the first week rather than at the end.
Questions

Before you ask for one.

How do you measure a saving honestly?

We agree the number and its source in writing before building: usually a count and a duration out of a system you already run, not a survey. We measure the same thing afterwards with the same query. If the measurement turns out to be arguable, we do not get to invent a better one later.

What if it saves less than we hoped?

Then we say so, and we would rather find that out from the second measurement than leave you to discover it. You are not carrying the risk of an optimistic estimate.

What happens once it is running?

It runs while we run the system, and it covers hosting, monitoring and fixes. If you take operation in-house it stops. We will put the term in the agreement rather than leave it open.

Can you work with our existing systems?

Usually. Anything with an API, a database or an export is straightforward. Anything that only exists as a screen somebody clicks is possible but slower, and we will say so before quoting rather than after.

What happens to our data?

It stays in your accounts, hosted in the EU, under a processing agreement. It trains no external model, and nothing leaves your tenancy that you have not agreed to.

Tell us the loop.

Thirty minutes with both founders. Describe the thing your people do over and over, and we will tell you what it is worth, whether it is worth automating, and what we would build.

Book the call