Client results · AU · NZ · UK

Automation that pays for itself.

Two real projects, start to finish: what the business was doing by hand, what we built, how long it took and what changed afterwards. No vague claims, just the numbers.

Client names are withheld under confidentiality agreements. Results depend on the project.

By the numbers

What these two projects actually changed.

94%

Straight through

Invoices the first client now processes with no human input at all.

74%

Fewer calls

How much quieter the phones got at the second client's three reception desks.

11 days

Fastest build

From the first call to a system processing real invoices.

38%

Fewer no-shows

Missed appointments the second client stopped losing chairs to.

Everything at once

The results and the words behind them.

Figures and quotes from the two projects below, plus other work with wholesalers, clinics, agencies and finance teams across Australia, New Zealand and the UK.

2 days

of manual accounts payable work removed every week

Building materials wholesaler · Sydney

Monday mornings used to be a nightmare. Now I check the dashboard over coffee and everything’s already done.
Accounts Manager · Sydney

94%

of invoices processed with zero human input

Invoice automation

We had a VA doing nothing but data entry. They automated the entire workflow in 12 days, and we have not needed to replace her since.

A full role’s worth of work, gone

Operations Director · AU

Case study 01 · Invoice Processing & Accounts Payable Automation

Building Materials Wholesaler

Building Materials Wholesale · Sydney, Australia · 12 employees

Delivered in

11 days

From two days of manual work a week to none at all

What it looked like before

A 12-person building materials wholesaler in Sydney was processing 200–300 supplier invoices a month entirely by hand. Their accounts manager spent two full days every week opening PDFs, typing figures into Xero, matching them against purchase orders in a spreadsheet and chasing approvals by email. It was slow, easy to get wrong, and it all depended on one person.

Three invoices had been paid twice in twelve months because of typing errors, costing them real money and hours of reconciliation.

What we built

  • 1Built a pipeline that reads incoming PDF invoices from their shared inbox and pulls out the supplier, invoice number, line items, totals and GST
  • 2Connected it to Xero, so invoices are created in draft and matched against existing purchase orders automatically
  • 3Moved approvals into Slack: the invoice details arrive as a message with Approve or Query buttons, so no more email chains
  • 4Anything that doesn't match a purchase order is held back and flagged for a person to look at
  • 5A dashboard showing processing status, the approval queue and monthly volume

Built with

PythonGoogle Cloud VisionXero APISlack APIPostgreSQL

What changed

Two days a week back

Two hours a week on accounts payable instead of two full days

No duplicate payments

Duplicate detection ended a recurring, expensive error

94% straight through

94 invoices in 100 are processed with no human input at all

Live in 11 days

Processing real invoices within two weeks of kickoff

In their words

“I genuinely didn't believe it would work this well. Monday mornings used to be a nightmare. Now I check the dashboard over coffee and everything's already done.”

Accounts Manager — Building Materials Wholesaler, Sydney (name withheld under NDA)

Case study 02 · AI Patient Chatbot + Appointment Automation

Multi-Location Dental Group

Healthcare / Dental Practices · Manchester, UK · 3 locations, 22 staff

Delivered in

14 days

74% fewer calls to reception, and no-shows down 38%

What it looked like before

A dental group running three locations in the UK had reception desks that were drowning. Each site was fielding 60–80 calls a day, mostly bookings, opening hours and reschedule requests. Receptionists had no time for the patients standing in front of them.

No-shows were running at 18%. Every missed appointment meant lost revenue and an empty chair. Reminder calls were meant to happen but were often skipped because everyone was too busy.

What we built

  • 1An AI chatbot on all three clinic websites handling bookings, reschedules, common questions and new patient intake
  • 2Connected straight to their practice management system, so anything the bot books appears in the calendar in real time
  • 3An automatic reminder sequence: text 48 hours before, email 24 hours before, text 2 hours before, each with a one-tap confirm or reschedule link
  • 4New patient intake collected by the chatbot before the first visit and filled into the practice system
  • 5Out of hours, the bot handles everything and escalates genuine dental emergencies to the on-call number

Built with

Next.jsClaude APIDentally APITwilioSendGrid

What changed

74% fewer calls

Receptionists now handle the complicated queries only

No-shows down 38%

From 18% to 11.2% across all three locations

Reception time freed up

Recovered from fewer no-shows and far fewer routine calls

4.8★ patient rating

Up from 4.1 in the post-visit survey within 60 days

In their words

“The phones still ring, but it's actually important calls now. The bot handles everything else. Our receptionists finally have breathing room, and the patients love being able to book at 11pm.”

Practice Director — Multi-Location Dental Group, UK (name withheld under NDA)

Your turn

What would your case study say?

Book a free 30-minute audit. We'll find the job that's costing you most to do by hand and give you a fixed quote to take it off your plate.