KAIRONULL
EU AI Act: Article 50 transparency live 2 August 2026  ·  Annex III high-risk (Article 12, Article 9) fixed at 2 December 2027 See obligations →

N° 0026 · Deployment Models

How KairoNull is deployed.

Three engagement models, structured around where your organisation is in its AI governance journey. Every model starts with a scoping conversation. No commitment required before that.

Model 01

Evidence Assessment

A fixed-scope assessment of one to three AI workflows. It shows what the workflow records today, what could be altered or lost without anyone noticing, and what it would need to record. It uses only redacted artifacts you supply; we do not connect to your systems or hold your data.

Includes

Fifteen evidence capabilities rated, each tied to an artifact you supplied
One reconstruction test of a real past decision
A target evidence design for that workflow
Prioritised gaps, each sized small, medium or large
Written report and a 45-minute findings call. Not included: legal or regulatory opinion, or any mapping to a law or standard
Request a scoping conversation →

Model 03

Enterprise Governance Infrastructure

For organisations with several AI workflows or business units. Scoped individually: nothing in this model is available off the shelf, and terms, hosting location and support are agreed in writing before any commitment.

Includes

Everything in Standard Deployment
Event design across several workflows
Deployment across business units, scoped per engagement
Hosting location and data-handling terms agreed in the contract
Request a scoping conversation →

Every deployment starts the same way.

A 30-minute scoping conversation. We discuss the AI workflow, what you record today, and which model fits. We do not give legal or regulatory advice. No commitment required. No sales process. If KairoNull is not the right fit, we will tell you that in the conversation.

Common questions

How long does deployment take?

Timelines are scoped per engagement and are not promised here. A first test record can be sent in a few lines with the Python or Node.js SDK (see the quickstart); a real workflow takes longer and depends on the system.

Do we need to change our AI models or infrastructure?

Not your models. Your integration sends events to KairoNull, so some code in your workflow changes; how much depends on the workflow and is part of scoping.

Where is the evidence stored?

Today the evidence ledger is hosted by KairoNull in one EU Supabase project, and the events you send are stored there. Hosting inside your own environment (on-premises or private cloud) is not a standard service; it would be a separate project, scoped and agreed in writing before any commitment.

Can the evidence be verified without KairoNull?

Yes, for integrity. Record hashes, chain links and the RFC 3161 timestamp token can be checked offline on an exported bundle with the independent verifier and OpenSSL, without calling KairoNull systems. The timestamp check needs the timestamp authority\'s own certificate, which you obtain from the authority. See the full methodology at kaironull.com/how-verification-works.

What regulatory frameworks does KairoNull address?

KairoNull records evidence about AI decisions. It does not certify or assess compliance with any law or standard, does not map its records to specific articles, and gives no legal advice. Whether a record meets a given obligation depends on your system and needs a legal assessment.