Skip to content

About

Make AI’s financial output verifiable. Make every check auditable.

AI is answering financial questions faster than anyone is checking the answers. Models quote filings and summarize earnings inside products used for real money decisions, and almost nobody traces that output back to a source. Artizar exists to close that gap.

The stance

Verifiability, not infallibility

No one can promise a model that is always right. We promise something you can hold us to: every claim checkable, every check on the record. Three commitments make that concrete.

  • Verdict

    A verdict is never an opinion.

    Confidence tiers come from a fixed rule table applied to structured evidence. The model gathers; it does not decide. The same claim, checked twice, returns the same verdict byte for byte.

  • Record

    The record outlives the run.

    Every verification writes a signed, append-only audit record. Months later you can show exactly what was checked and which rule produced the verdict, even after the model or the prompt has changed.

  • Evidence

    Evidence ends at a primary source.

    Every verdict traces back through an unbroken provenance chain to the filing itself, in EDGAR or PACER, not to a paraphrase of it.

Founders

The founders

  • Santiago Lozano

    Co-founder

    Santiago studied financial economics at Columbia and worked on M&A at CEMEX. He leads the financial side of the product: deciding which sources count as canonical and what a defensible verdict looks like.

  • Kevin Haddad

    Co-founder

    Bio coming soon.

Early, on purpose

The engine is in build and the product is pre-launch. We’re scoping the first integrations with a small group of early partners. If you’re putting AI in front of financial decisions, we’d like to talk.

Request early access

Artizar

Basque · the morning star