Ensign · how it works

Proprietary instrument

ENSIGN

An ensign is the flag that shows where a ship actually comes from.

A label feels like it solves the problem. It doesn't, on its own.

Most agencies think a small "AI-generated" tag is the safe move. The ASA has been explicit: a disclosure doesn't rescue a claim that's misleading regardless of how it was made.

An agent, not a questionnaire

Built on your world, not an invented one

The usual way

Tests whether people know the labelling rules.

The Ensign agent

Checks the actual content and client clauses against the rules.

Step one

What it works with

Content workflowClient contractsApproval recordsThe markets each asset runs in
Ensign agent

Opt-in and anonymised where people are involved.

Step two

How it investigates

connectcontent workflow and client contracts: read-only
check48 assets this week across 5 markets
match3 need an AI label under the rules of the markets they run in
conflicta client clause bans AI imagery. Asset 12 uses it.
testasset 31 is labelled, but its claim is misleading regardless
recordlabel, reason and approver stored for every asset

Every line is something a person can check afterwards.

Step three

What it produces

AssetResultReason
Asset 12ConflictClient clause
Asset 31MisleadingA label does not fix it
Asset 07Label neededEU markets
Asset 19ClearRecorded

A content review with a record for every asset.

How it is held to account

Rules, not vibes

01

A label does not rescue a misleading claim

The claim is tested on its own, whatever the tag says.

02

The market decides the rule

Each asset is checked against the rules of every market it will run in.

03

Every decision is recorded

What was labelled, why, and who approved it.

The agent is held to the same questions it asks of everyone else.

Configured to your regulator

The same agent. Your market.

ASA guidance and the CAP Code on advertising that uses AI.
Swiss advertising and data protection rules for AI-made content.
The EU AI Act's transparency duties for AI-generated content.

Choose a market. Other markets are added as packs.

How an engagement runs

From first conversation to a working agent

1

Scope

Which clients, which markets, which asset types.

2

Connect

Read-only connections to your workflow, reviewed first.

3

First review

Every asset checked, every conflict listed.

4

Keep a record

The evidence for each decision, ready to show.

A typical shape. The sequence and timings are set with you.

One platform

Ensign does not work alone

KeelSets the policy and the gates for every use case.
SentinelQuestions what each agent can see, change and cause.
SoundingsSees what is actually running across the organisation.
BearingProves it to the supervisor, obligation by obligation.
The human layerHALO, Squall and Ensign measure what no system log can.

See the whole system on the Solutions page.

That's Ensign

Find out whether your disclosure judgement would actually hold up.

A demonstration of the method is one click away. The deployed Ensign is built for your environment and configured to your regulator.

Discuss a deployment