What you've declared, one page.
The five doors AI enters a company through — sanctioned, vendor-embedded, shadow, agentic and inherited — recorded as your organization declares them, not as a scan discovers them.
LoadLine™ — the standing instrument
A load line — the Plimsoll line — is the mark on a ship's hull that makes the vessel's load visible against a stated line. LoadLine draws that mark across the AI systems and answers an organization declares. It is an instrument that supports an expert-led AI risk assessment through operator-entered inventory, role-routed questions, captured answers, and decision records leaders can examine.
§ 01 — What it shows you
The five doors AI enters a company through — sanctioned, vendor-embedded, shadow, agentic and inherited — recorded as your organization declares them, not as a scan discovers them.
In an expert-led assessment, LoadLine records declared systems and captured answers, can link supplied evidence records, and supports operator-created pathways and decision records. Source-document custody, independent verification, and production self-service are not claimed.
A consultant's report is true the day it is signed. LoadLine is a standing record your team can update as new evidence is entered, so what you are carrying stays examinable rather than frozen.
§ 02 — The Five Doors
AI enters a company through five doors. Most inventories watch one. Answer five questions — free, two minutes — and watch your own load become visible.
Your declared estate
Every organisation draws this somewhere. LoadLine is where you write down where.
Sanctioned AI
Do you have a current list of the AI systems your organisation approved and deployed?
Your answers stay in your browser. Nothing is sent or saved.
§ 03 — Watch
§ 04 — Where LoadLine sits
Ask two questions of any AI risk tool: how long does its answer stay true, and where in the failure does it sit?
A standing, pre-impact instrument. It makes declared exposure examinable while risk is still a pathway rather than a loss — so leadership can decide before impact, not explain after it. This is where LoadLine sits.
Small error
Across a step
Across a seam
Across systems
Surfaces as loss
§ 05 — What the Baseline actually does
Record the estate — all five doors, as you declare them. Most leaders find the act of writing it down is itself the surprise.
Trace the pathways — working with our assessors, your team maps the routes a failure could travel and the seams where context is lost.
Ranked against your tolerance — you define the line during the assessment; each pathway measures against it by business impact and how far its failure chain has progressed, so the riskiest findings order to the top. Views stay available by pathway, by system and by business unit; the decisions are your leaders' to apply.
Keep it standing — as your team enters new evidence, the record updates, so the picture your leaders examine is the current one.
Evidence everything — findings that are sourced, dated and replayable, so the reasoning behind a decision can be examined later.
§ 06 — How it's bought, what you keep
Fixed fee, fixed scope, about three weeks — and 100% credited toward a subscription if you continue. Success criteria you write. You keep the scorecard, the findings briefing, and the fix list either way. Exact numbers in the first conversation.
The instrument that stays. Founding-cohort terms for early customers who help shape it.
Nothing to install
LoadLine is our instrument, run with our assessors. No software lands in your environment.
No system access required
It works from what your organization declares and the evidence you choose to supply — no credentials, no connectors, no agents.
Not an IT project
The Baseline takes your leaders' time, not your infrastructure. IT is welcome in the room; it isn't a dependency.
§ 07 — Who asks for it
who want the answer to 'how are we managing AI risk?' to take an afternoon, not a discovery fight.
who suspect AI losses are already in the P&L under other names.
tired of being accountable for doors nobody instruments.
who just inherited an AI estate nobody diligenced.
§ — A private briefing
We're early, and we say so — no customer logos, a working instrument you can see in a briefing, and a proof standard we hold ourselves to.