Soundings · how it works

Proprietary framework

SOUNDINGS

Knowing what's running, and whether it's watched properly, before you build anything.

Nobody can answer a simple question

Ask five people in the same company whether AI is being used correctly, and you get five different answers. Nobody agreed what correct means.

An agent, not a questionnaire

What it reads, not what people remember

The usual way

Asks people which AI they use, and relies on them to remember and to say.

The Soundings agent

Reads the spend, sign-in and software records that show what is actually in use.

Step one

What it connects to, read-only

Finance and expensesIdentity and sign-inSoftware estateCloud accountsNetwork logs
Soundings agent

Read access only. Nothing is changed in your systems.

Step two

How it investigates

connectidentity, finance and software records: read-only
discover61 AI tools found. 38 are on the approved list. 23 are not.
tracecomplaint-triage assistant touches customer data and decides which complaints escalate
scoreregulation high, data high, reversibility low. Overall high: the highest of the three, never an average.
compareleadership believes 38 tools are in use. The records show 61.
flag4 new tools since the last run. 1 touches customer data.

Every line is something a person can check afterwards.

Step three

What it produces

Use of AIRatingOversight
Complaint triage assistantHighGap
CV screening modelHighGap
Meeting summariserMediumMatches
Marketing copy draftingLowMatches

A live AI inventory with the oversight gaps flagged.

How it is held to account

Rules, not vibes

01

Three scores, highest wins

Regulation, data and reversibility are scored separately. The overall rating is the highest, never an average.

02

Records beat opinions

A tool that appears in the records is on the map, whoever declared it.

03

New tools raise a flag

Each run is compared with the last. Anything new is flagged.

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

Configured to your regulator

The same agent. Your market.

Consumer Duty, UK GDPR and FCA expectations decide which uses carry high regulatory exposure.
FINMA guidance and the Federal Act on Data Protection set the regulation and data scores.
The EU AI Act risk classes set the regulation score. GDPR sets the data score.

Choose a market. Other markets are added as packs.

How an engagement runs

From first conversation to a working agent

1

Scope

Which records, which business units, who owns the answer.

2

Connect

Read-only connections, reviewed by your security team first.

3

First map

The inventory and the gaps, each linked to its evidence.

4

Keep watching

The agent re-runs and flags what is new.

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

One platform

Soundings 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 Soundings

Want to know what's actually running in your organisation?

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

Discuss a deployment