Tax technology built through the Canadian lens
We build AI products that read tax work through the Canadian lens: the Act, the regulations and CRA guidance as the model itself, not a setting applied at the end. They give professionals their hours back, keep client data sovereign, and hold to a standard that competes anywhere in the world.
give Canadian tax professionals back the hours that software should never have taken from them.
keep the country’s most sensitive financial records sovereign, Canadian by architecture and not by policy.
prove that Canadian-built tax AI can set the standard the rest of the world measures against.
Three commitments, and everything follows from them
Whatever we build, and in whatever order we build it, it has to do these three things. If a feature doesn't, it doesn't ship.
Give back the hours
Mechanical work, whether re-keying, reconciling or chasing the same answer through three subscriptions, is work software should absorb. We take it, so the time returns to judgment, advice and clients.
Hold the line on security
Tax records are among the most sensitive data a person or company has. They stay encrypted, scoped to the engagement, resident in Canada, and under Canadian jurisdiction. That is architecture, not a policy page.
Canadian first, world class
Built here, for the Canadian rulebook, by people accountable to it, and engineered to a standard that stands beside anything built anywhere else. Local depth is our advantage, not our ceiling.
Canadian tax law isn't a setting. It's the model.
Most tax software localises: a foreign engine with Canadian labels bolted on, reconciled by hand at year-end. We invert that. The statute, the regulations and CRA guidance are the structure our systems reason in, so what comes out is already defensible here.
Reasoned in the statute
Our models work from the Canadian rulebook as their frame of reference, not a generic ledger with a currency field.
Sourced, every time
Every conclusion names the provision, guidance or decision it rests on. An uncited output is a defect, not an acceptable answer.
A professional decides
Automation prepares and evidences. A named person approves, and that approval is on the record for as long as the file lives.
Security and sovereignty, decided in the architecture
Where data lives, who can compel it, and whether a decision can be reconstructed: these are engineering choices made before the first feature, not commitments added after.
Resident in Canada
Built and hosted on Canadian infrastructure. Client records stay under Canadian jurisdiction, with no cross-border processing.
Scoped to the engagement
Encrypted in transit and at rest, with access bounded per engagement, so professional and privilege boundaries hold inside the software too.
Citation as a requirement
No automated conclusion ships without its source. We treat an unsourced answer as a bug in the system, not a limitation of AI.
Reconstructable history
Every suggestion, approval and edit logged against a named user, so any position taken can be explained years later.
Depth in one jurisdiction is how you earn the rest
Tax is the most jurisdiction-specific software there is, which is exactly why the global players go shallow everywhere. We went deep here first.
The hard part isn't the interface. It's building systems that reason inside a statutory framework and can show their work. Solve that for Canada and the method travels. The rulebook changes; the discipline doesn't.
Canadian-founded, Canadian-hosted, accountable to the profession we serve.
Engineered to the standard of the best software in any category, not the best in ours.
Statutory reasoning that shows its authority is portable. The rulebook is what changes.
That the global benchmark for tax AI is a Canadian one.
Hear it from the founders
Two minutes on why Canadian tax practice needed its own infrastructure, what we will not compromise on, and when it opens. We are deliberately not publishing product screens: software you'll trust with client files deserves a real conversation, against a file that resembles yours.
Early-access firms are invited in order, keep founding terms, and shape what we build next.
Questions we get
most often
Anything else, ask us directly and we answer every message.
Contact the teamWhat does TaxClever actually do?+
We build AI products for Canadian tax work that reason inside the Canadian rulebook, cite the authority behind every conclusion, and keep client data resident in Canada. They are in private beta with a small group of firms, and we show them in a guided session rather than publishing screens.
What does "through the Canadian lens" mean in practice?+
That the Act, the regulations and CRA guidance are the structure our systems reason in, not labels applied to a foreign engine after the fact. It is the difference between software that produces defensible Canadian output and software a professional has to reconcile by hand.
How do you keep AI output defensible?+
Every automated conclusion carries the provision, guidance or decision it rests on, and every suggestion, approval and edit is logged against a named user. An uncited answer is treated as a defect in the system.
Where is client data stored?+
In Canada, on Canadian infrastructure, encrypted in transit and at rest with access scoped per engagement. There is no cross-border processing, and no foreign jurisdiction over your clients’ records.
Why don’t you show the products on the website?+
Because a marketing reel is a poor basis for trusting software with client files. We would rather show the working product to a professional who can interrogate it, against a file that resembles theirs.
Canadian first: does that limit you?+
The opposite. Tax is the most jurisdiction-specific software there is, which is why global players stay shallow everywhere. Solving statutory reasoning properly for Canada builds a method that travels; the rulebook is what changes, not the discipline.
See it against your own file
Get early access and we will walk you through the platform on a file that resembles yours. Early-access firms keep founding terms.