Case Studies

Multilingual customers, one shared view

Enquiries in several languages across five channels, visible to the whole team.

Client: A private education and training provider in New Zealand
Business background: The provider serves students and families from different countries and cultural backgrounds, communicating in English, Chinese and other languages. Course enquiries, enrolment, payment and follow-up arrive through the website, email, phone and social messaging channels.
Operational challenge: Customer information sat in individual inboxes, spreadsheets and chat histories. It was hard to see which market a customer came from, which language they preferred, what courses they had asked about, whether a quote had gone out, who was following up and when. Enrolment rarely completes in one conversation, so enquiries were missed, customers were contacted twice, and history was lost when staff went on leave or moved on.
What we reviewed: The full journey from first enquiry to enrolment and beyond: how enquiries arrive on each channel, how the team learns a customer's background, language preference and learning needs, how course material, quotes and enrolment requirements are sent, and how follow-up runs through payment, onboarding and after the course.
Solution: One customer management system holding contact details, market of origin, preferred language, courses of interest, enquiry source, enrolment stage, communication history, required documents and payment status. Multilingual communication was designed deliberately rather than handed to auto-translation: approved templates carry routine messages, while anything covering price, refund policy or enrolment conditions is reviewed by a person so every language version says the same thing. Replies are prepared in the customer's language with an English summary alongside — so staff who speak the customer's language are no longer the only people who can serve them.
Implementation: Follow-up reminders were built around each enrolment stage — enquiry not yet answered, course information sent but no appointment, quote sent but unconfirmed, documents incomplete, enrolled but unpaid, course starting, and a check-in after the course. Management reporting was added covering enquiries by language and market, channel performance, course interest, where customers stop, and conversion from first enquiry to enrolment.
Training and support: The team was trained on the shared workflow, and the privacy settings were documented in plain language around the Privacy Act 2020: what each piece of information is collected for, who can access it, how long it is kept, and how access and correction requests are handled — with privacy notices customers can actually understand in their own language.
Business result: Customers no longer repeat their story to each new person. The team serves customers across languages together, enrolment progress no longer depends on one person's memory, handovers survive leave and role changes, and management sees demand by market and channel instead of guessing.

Manufacturing on one system

Paper dockets and spreadsheets replaced by one system — live in two months.

Client: A manufacturing business
Business background: The company produces a wide range of industrial products, with more than ten thousand product variants across its catalogue. Production had grown steadily while day-to-day records stayed on paper.
Operational challenge: The factory ran on paper job dockets and spreadsheets. Materials, stock, job numbers and production progress were tracked by hand, so the true state of production was difficult to see.
What we reviewed: How an order moved through the factory from intake to dispatch, where information was recorded along the way, and which records existed only on paper or in individual spreadsheets.
Solution: One system covering materials, stock, job numbers, production steps and output, replacing the paper dockets and parallel spreadsheets.
Implementation: The system was configured around the factory's existing workflow and brought into production use within two months.
Training and support: Staff were trained on the new system as part of the rollout, and it has been supported and refined in ongoing use.
Business result: Live in two months, and in daily use since. Production now runs on one system instead of paper dockets and spreadsheets.

AI inside a private network

Models and data kept entirely inside the client's own network.

Client: A business client in New Zealand with strict data requirements
Business background: The organisation held sensitive information that was not permitted to leave its own network, which ruled out cloud-based AI services.
Operational challenge: The team wanted to search and question its own material with AI, without any data passing through an external provider.
What we reviewed: The client's servers and network environment, the material the knowledge base would cover, and the security requirements the solution had to meet.
Solution: A knowledge base built entirely on the client's own servers, with both the models and the information staying inside their network.
Implementation: The models and knowledge base were deployed on the client's own hardware, with nothing sent outside the network.
Training and support: The client's team was shown how to use the knowledge base and how to keep its content up to date.
Business result: AI over sensitive information, with the models and the data staying inside the client's own network.

A moving company, one customer record

One customer record to replace four enquiry channels and a paper notebook.

Client: An owner-operated moving company in Auckland
Business background: A moving business serving the Auckland market at the mid to upper end, running a small fleet and a full-time crew. Work comes in through social media, search and a third-party moving platform that takes a commission on each job won.
Operational challenge: Enquiries arrived through four separate channels. Quotes were tracked in a paper notebook and in chat history, so there was no reliable record of what had been quoted to whom, no visibility of which enquiries had gone quiet, and no way to tell which advertising channel was actually producing paid work.
What we reviewed: The full process from first enquiry through quoting, booking, the job itself, payment and follow-up, in a working session with the owner. We mapped where information is created at each step, which systems already hold it, and the quality of the historical customer data.
Solution: A CRM as the single customer and commercial record, covering lead capture, job qualification, quoting, booking, job outcome and customer history. Deliberately bounded: the accounting system stays the financial source of truth, and payroll and dispatch stay outside the first phase rather than being absorbed into it.
Implementation: A three-month first phase: centralise every new enquiry, standardise the quote and booking fields, add follow-up tracking, migrate the clean historical customer data out of the accounting system, connect completed jobs back to it, and add reporting that attributes revenue to each advertising channel.
Training and support: Training and acceptance at the end of the first phase, then an ongoing managed service with monthly reporting on enquiries, conversion and channel performance, and a quarterly review.
What changes: One customer record from first enquiry to final invoice, follow-ups that surface instead of being remembered, and revenue attributed to the advertising channel that produced it.

None of these began as a whole-business project. Each started with one workflow.

Where we work

What this looks like in your industry

These are the industries we work in around Auckland and the problems we are asked about most often. If yours is not listed, the starting point is the same: we look at how the work actually flows before recommending anything.

Building and construction

Typically

Quotes are written up from site notes and photos after hours, variations are agreed verbally on site and forgotten by invoicing time, and the real cost of a job is only clear once it is finished.

What we put in place

One record per job from enquiry to final invoice, with variations captured when they are agreed and actual cost visible against the original quote.

Plumbing, electrical and trades

Typically

Work is booked by phone and text. Nobody knows which callers never received a quote, and maintenance that should come round again is only remembered when the customer rings.

What we put in place

Every enquiry captured, quotes tracked through to won or lost, and recurring work scheduled instead of remembered.

Gardening and grounds maintenance

Typically

Recurring rounds live in a diary or a spreadsheet. It is hard to see which properties are due, which were skipped in bad weather, and whether a site still pays for the time it takes.

What we put in place

A recurring schedule with visit history for each property, and time on site measured against what is charged for it.

Property management

Typically

Maintenance requests arrive by email, text and phone. Owner reporting is assembled by hand each month and compliance records sit in several places.

What we put in place

One record per property, requests tracked through to completion with cost attached, and owner reporting produced from the same record rather than rebuilt.

Clinics and practices

Typically

Reception answers the same questions all day. Gaps and no-shows are only visible once they happen, and recalls depend on someone remembering to make the call.

What we put in place

Routine enquiries answered from your own material, recalls and reminders scheduled rather than remembered, and gaps offered to a waiting list.

Hospitality and cafés

Typically

Customers find you through search and reviews, but the listing is out of date. Supplier invoices and stock are on paper, and it is unclear which items actually make money.

What we put in place

Local search presence and reviews kept current, and sales connected to supplier and stock cost so margin is visible per item.

Galleries and studios

Typically

Works, consignment terms with artists, and collector contacts sit across spreadsheets and email. There is no single view of what is where, or who bought what.

What we put in place

One catalogue of works with current status — available, consigned, on loan, sold — with artist terms and collector history attached.

Packaging and light manufacturing

Typically

Many product variants and specifications, with quoting that depends on the person who knows the catalogue. Stock and job status live in spreadsheets.

What we put in place

Structured product and pricing data so quotes are consistent whoever prepares them, with job and stock status visible in one place.

Business problems we've solved.

Multilingual customers, one shared view

Manufacturing on one system

AI inside a private network

A moving company, one customer record

What this looks like in your industry

Building and construction

Plumbing, electrical and trades

Gardening and grounds maintenance

Property management

Clinics and practices

Hospitality and cafés

Galleries and studios

Packaging and light manufacturing

An AI assistant for the enquiries you answer over and over

Tell us what is making your business difficult to run.