01 · MODERNISING OLDER SYSTEMS
We upgrade software your business depends on — without stopping the business to do it.
Your core systems weren’t built for AI. We fix that.
We modernise the software organisations already run on — upgrading systems that have been in service for years, wiring AI into the places it actually pays off, and rebuilding what genuinely needs rebuilding. For governments, listed companies, and funded startups across Canada and the US.
Most AI projects don’t fail at the model. They fail at the seam between the model and the decade-old system that has to use it. That seam is where we work.
CLIENTS HAVE INCLUDED THE GOVERNMENT OF ONTARIO · A US-LISTED PUBLIC COMPANY · CANADIAN STARTUPS References available on request.
02 · ADDING AI TO WHAT YOU HAVE
AI wired into real work — with the cost controls that keep it useful a year later.
03 · REBUILDING FOR AI
When patching stops paying, we rebuild properly — and say so plainly when a rebuild is the wrong call.
04 · KEEPING IT RUNNING
We stay on afterwards, as the platforms and services underneath keep changing.
THE RECORD
Anyone can claim a track record.
Here is ours, with the paperwork.
Wiring AI into a system starts with understanding how that system survives change — and we have run machine learning inside one, in production, for three years. Each of these has a document behind it — a certificate, a signed agreement, a release history. Ask and we will walk you through them.
Three years inside a government justice system
Living Water Tech was engaged as a supplier on the systems behind a provincial government’s courts, contracted the way public-sector work actually happens — through a vendor of record. The engagement ran close to three years and outlasted a change of contract holder: the paperwork around the work changed hands, the work did not stop.
Government systems are where “we’ll rewrite it properly later” goes to die. They carry obligations set in law, they cannot go dark for a weekend, and the people depending on them never chose them. Learning to change software under those conditions is what this company was built on.
Living Water Tech Inc. is incorporated
Federally incorporated in Canada under the Canada Business Corporations Act, registered in Ontario. Not a side project with a landing page — a company that files, invoices, and carries liability.
A multi-platform accounting SaaS
Engaged under a master services agreement: a native Android app, a web application, and one .NET backend serving both — plus payment reconciliation against Stripe and automated PDF reporting.
The part users actually feel is the capture: machine learning running on the phone itself finds the receipt in the camera frame, straightens it, and pulls the figures off it — so the person holding the phone types nothing. Getting that to work reliably on thousands of different Android devices, on paper that is creased, faded, or photographed badly, is the whole job.
The client is not named here. We do not publish client identities without written consent, and we would extend you the same discretion.
That same system is still shipping
Year after year: crash fixes, tax-rule changes, operating-system changes that break old code, extra languages added long after the original build, and an automated release process that keeps working while the services underneath it change.
On Google Play the Android edition shows 50,000+ installs and a 4.6-star average across 696 ratings, with updates still landing in June 2026. The client is not named here, but those numbers are public record — anyone can go and check them.
A long run of continuous releases doesn’t prove the product made money — that credit belongs to the client. It proves something narrower: the engineering survived, and kept working for the people using it.
OUR OWN SYSTEMS
What we build when the client is us
Built and run on our own account — this is where the standards above come from.
WhaleClues ↗
A subscription product for ETF traders. It puts two things on one chart that normally live apart: where market makers are forced to hedge, and where the largest trades are actually happening — so a trader can see which levels a price is likely to turn at, and whether real money is behind them.
We designed it, built it, and run it, including the parts that are not code: per-symbol pricing, free trials, billing, and knowing whether any of it is working. Building software for a client and owning a product that has to earn its own revenue teach different lessons; we have done both.
An AI interface for an existing parts system
A .NET service that takes a system already in use — distributor inventory, pricing, and datasheets for electronic components — and exposes it as tools an AI assistant can actually call, rather than guess at. Around it sits the part most demos skip: enterprise API-key management, per-call usage accounting, and an authentication mode that refuses to start at all rather than run unguarded outside development.
This is the seam from the top of this page, built end to end: a model on one side, a real system with real access control on the other.
A Bible study app for iPad
Built for one stubborn problem: getting through a semester of dense English theology when English is your second language. It turns 1,300 structured documents into a workspace where you can search across every book at once, read any sentence in both languages side by side, and lift a quotation with its page reference already attached — which is what actually stands between a reader and an argument they can defend.
The shape travels. Any field that lives inside large bodies of specialist literature — legal, regulatory, medical — has this same problem underneath, and the same answer: make the corpus searchable, keep every claim traceable to its source.
An engineering system for AI agents
Our answer to a question many technical teams are now asking: how do you get real output from AI coding agents without quietly losing control of quality, cost, or your own engineers' judgement? It encodes verification rules that each came out of a specific production failure — a test suite that silently stopped running while its gate stayed green; an empty query result mistaken for absent data; a change reported as shipped whose pull request had never merged.
HOW WE WORK
Four commitments
VERIFIED
Done means proved
“Done” means we proved it works — not that it looked fine on our machine. A passing gate that never actually ran is worse than a failing one. Test gates assert expected counts; a deployment is proven layer by layer — merged, deployed, running — never inferred from one of them.
LEGIBLE
The reasoning ships with the code
Configuration and infrastructure carry comments explaining why they are the way they are, including what was tried and rejected. Whoever touches it next — often you, without us — inherits the reasoning.
CANDID
Straight answers about trade-offs
If a simpler approach would serve you better, we say so before the invoice, not after. Every recommendation carries what it costs you, not only what it buys.
DURABLE
We stay after launch
Software is not finished when it ships; that is when it starts accruing maintenance. We are set up for the multi-year part and we price for it honestly.
CONTACT
Tell us what you’re running.
A short description of the problem is enough to start. We read and answer every enquiry ourselves.
info@livingwaterapps.comLIVING WATER TECH INC. · INCORPORATED IN CANADA 2022 · ONTARIO