How Riskonami works

Riskonami runs a structured IRAM 2.0 assessment of a named system. You supply context and judgement. The server owns the method. Optional AI accelerates drafting — it does not replace scores, UNKNOWN, or sign-off.

What you walk away with

One assessment produces six named downloads from a single snapshot: executive, engineering, a control register that names vendor-held controls, overall risk and overall compliance (these include accepted vendor answers), and a vendor control assessment per named supplier. That is the artefact you commission — not a chat transcript.

Sample reports How vendor assessment works

Nine phases, one engine

Context and impact → architecture and scope → control points → threats → control strength → risk scores → your decisions → treatment → issued report. The same engine will later support unattended and voice input. Today the live product is the guided (phased) walk.

Riskonami v2 assessment spine: phases P1 through P9, optional named AI calls in red, server work in gray — Blue is your work. Red is a named AI call (1 token if it runs). Yellow is a Yes/No or button. Gray is server work and is not billed. There is no Phase 10.

AI accelerates. Humans remain accountable.

You can run the whole assessment with AI off. When AI is on, Riskonami uses a handful of named calls (context from your files, diagram reading, control proposals, optional narrative). The model cannot invent phases, set residual scores, or formally accept risk.

Assurance — what AI does and does not do AI in the product

Free process, paid enrichment

Working the phases by hand is free. Each named AI call that actually runs costs one token (typically two or three for a full assisted walk). Download your reports — we do not keep server copies forever.

Pricing Tokens in plain language

Ready to try it, or still deciding?

How-to lives in the docs. This page is the decision brief.

Product documentation Talk to us