Every filing out correct and on time — without re-checking everything yourself

For the people who run customs declarations, IP renewals, fund administration, and KYC packs. The work arrives as files and leaves as filings, and every error costs real money.

The questions you'd ask first

Who is liable when AI gets a filing wrong?

You are — same as today. We design for that instead of pretending otherwise. Your team keeps final say on every document: we build the draft step, never the send step. Nothing reaches a regulator, a client, or a registry without a named person signing off. And every prompt and every output is logged, so when someone asks "how was this filing produced?", there is a written answer.

Does client data leave the EU?

No. What we build is designed for EU data residency. And the audit itself needs none of your data to start: anonymized samples, or just a conversation about how the work flows, is enough. No system access required.

What if the AI invents something into a regulated document?

That risk is real, which is exactly why the human check never comes out. We build the assembly work around your judgment, not a replacement for it: the AI drafts from your source documents and flags what it could not find. It does not get to decide. A reviewer sees every draft next to its sources before anything is filed.

What this usually looks like

Genre-level sketch — our guess about firms like yours, not a claim about your firm. Documents arrive by email. Someone assembles the dossier against a checklist. Deadlines live in a spreadsheet and in somebody's head. A senior person re-checks everything, because an error is a liability, so peak load means evenings.

Our guess: the assembly and the deadline-tracking are the expensive part. Not the judgment. We're set up to build exactly that layer — dossiers assembled from incoming files, gaps flagged, deadlines surfaced — with your reviewer's sign-off as the one step the system cannot skip.

The free launch audit

While launch slots last, workflow audits are free in exchange for honest written feedback. One 30-minute call. Your three best AI opportunities, ranked, in writing. No system access, no preparation, no sensitive data. We're a new firm: the audit is how we earn the right to build anything.

We review everything that ships. This page was drafted with our own AI stack and reviewed by a human before it shipped. That review discipline is the product.

One line is enough. Email rachel@orellis.ai, or leave one line below — we'll reply with two time options.