The best engagements start when we understand three things: what's your pain point, how the answer
needs to land inside your organisation, and what we're allowed to see.
Everything is optional. A sentence is plenty. Click the options that fit, type only where you want
to. Your answers save in this browser as you go, when you're done, one click sends them back to us.
01 · The problem
What's your pain point
Not the risk category but the actual decision.
What did you learn, when, and when did the information actually first exist?
Select what you already use, then tell us the question it can't answer.
If we did one thing that made this worth it, what would it be?
02 · The deliverable
How it needs to land
AnalystsGeneral counselExec
committeeBoardA system (API / data feed)
Written assessmentStanding briefingAlert feedStructured data into our systemNot sure. Advise us
One-off deep diveWeeklyMonthlyEvent-driven onlyContinuous
Flag everything materialOnly above a threshold we
defineLet's work this out together
This determines how we document sourcing and confidence levels.
Internal use onlyAuditRegulator
reviewLitigation discoveryBoard scrutiny
03 · Data access
What we can see
Our work is built on open sources and our own networks. It functions
with zero access to your internal data. It sharpens considerably when we can see how risk affects your
operations. Pick the level you're comfortable with today; you can move at any time, and nothing is connected
before a security review with your IT team.
Nothing. Public sources only Level 0
Deliverables are specific to your sector, generic to your exposure.
Curated share Level 1
You place selected documents in a secure, access-logged data room. We read them, and nothing is copied into
any
model or shared system. Verified deletion on close.
Retrieval in your environment Level 2
Using RAG (Retrieval-Augmented Generation), we run a retrieval layer over your document stores, deployed
inside your own tenant or private cloud.
Documents and embeddings never leave your environment, our analysis queries them at run time, mirroring your
own access permissions, fully audit-logged, nothing persisted on our side.
Standing integration Level 3
Level 2, plus our structured output flowing back into your systems via API, so our signal and your internal
data enrich each other continuously.
Your data and our models. By default your data is never used to train or improve any
model, ours or a third party's. It is used only to answer your questions, in your environment. Separately, and
only with explicit written opt-in, we may use derived, anonymised signals to improve our detection models,
never raw documents, never anything attributable to you, never shared across clients. Saying no changes nothing
about what you receive. Standard at every level above 0: NDA and DPA before access, encryption in transit and at
rest, EU data residency where required, named individuals only, access logs on request.
04 · Anything else
We'll respond within a week with a scoped proposal: the specific problem we'd attack first, the
deliverable format, the access level it assumes, and a fixed four-week timeline for the first proof.