Avogado is built around how legal work actually happens, not around a chat box. Eight operating principles separate it from a general model that will answer anything, confidently, whether or not it’s right.
Avogado will not cite a case, statute, regulation, or ruling it hasn’t actually retrieved and read this session. Drafting language can come from the model; legal substance (specific authority, current law, a regulator’s position) must be verified against the primary source. Tax and financial-services questions, where fabrication risk is highest, go to the authority’s own site.
Each statement is tagged by how Avogado knows it: fetched from a primary authority, drawn from an indexed legal corpus, taken from your own input, or reasoned. At the end, those tags aggregate into a plain confidence picture.
Instead of five clarifying questions or a vague “it depends,” Avogado asks the single question that changes the answer, makes its best inference on the rest, and flags what it assumed. When your instruction is wrong, risky, or below market, it says so in one line before complying. It takes a position; it does not deflect to “consult a lawyer.”
Each matter is its own workspace: its own documents, drafts, research, and jurisdiction. Nothing leaks between them. Every consequential decision is written into a durable, challengeable audit trail, so you can come back months later, ask why a clause reads the way it does, and get the actual reasoning: the candidates considered, the call made, the authority behind it.
Drafting runs in stages: find real exemplars from the corpora, compare them dimension by dimension and decide what to anchor on, then draft in your voice, partisan to your side. Two independent passes follow before anything reaches you: a style red-team that strips the tells of AI drafting, and a citation-verifier that re-checks every source. Every draft ships with an annotated companion: the reviewing lawyer’s copy, showing the work behind each clause.
A diversified shortlist of real clauses.
Weigh, find market-standard, record why.
In your voice, then self-checked.
Avogado drafts from an indexed library of real legal material: model contracts, attorney-annotated agreements, soft-law principles, government model forms, and statute & case-law corpora, searched with hybrid retrieval that blends keyword and meaning so the right clause surfaces from the right document. New corpora ship continuously, and live web-fetch covers primary authority when a question turns on current law.
Built for cross-border practice (Japan, the US, UK, EU, Singapore, Hong Kong, Korea, and Colombia among them), it reads source documents in the language they’re written in and replies in the language you use. When a matter touches a new jurisdiction, the relevant corpus loads on demand.
Avogado writes its own task list the moment you give it work, and keeps a running record of open items, pending decisions, and to-dos. When something needs you, it surfaces front and center; when it’s done, it’s checked off with a time. Nothing waits silently in a chat scroll.