Not a hard question. A simple one, asked by boards every quarter: what is the actual state of our data? Three decades of advisory work kept producing the same answer — nobody could say.
A capable data team. Tooling that had been properly selected and properly paid for — a catalogue, a quality engine, an orchestrator, a warehouse. And a leadership group that still could not describe the state of its own data without commissioning somebody to go and find out.
The organisations with the worst visibility were rarely the ones with the least technology. Several had bought more than they could use. Each tool worked; each answered for its own layer, and none answered for the whole.
The teams were good and usually knew where the problems were. What they could not do was express that knowledge as something a board could read, compare across domains, and track from one quarter to the next.
Not more data, more tools or more people. One measure of the organisation’s data that gives the same answer whoever is asked, and can be produced again next month without another project.
This is not a criticism of those categories. They do their jobs, and you should keep them. It is an observation about what nobody built.
Describe what exists. They do not measure whether it is any good, and they carry no lifecycle.
Watch pipelines and raise alerts. They watch the plumbing, not the product the business depends on.
Hold the control framework. They record that a control exists, not the condition of the data beneath it.
Produce a number once. Accurate the week it is written, and unrepeatable at any sensible cost.
One establishes where you stand. The other stops that answer going stale. Either works on its own, and neither requires the other.
The founder’s advisory practice, delivered as a service — senior data expertise at the cadence you need it, without a permanent hire.
Continuous measurement of every data product, computed from evidence rather than asked for in a survey.
No rebuild, no migration, and no requirement to standardise first. We read from the catalogue, the quality engine and the warehouse you already run.
Thirty days on a real slice, with success criteria agreed before anything begins — and you keep the evidence pack whatever you decide afterwards.
Every score lands on something a person owns, so it carries a name, a finding and a next move rather than a grade for the organisation.
The same measurement is read by the board, by risk, by governance and by the people who depend on the data — without anyone re-cutting a deck.
If your real constraint is funding, sequencing or ownership rather than measurement, that is the finding you get — from the same people who would otherwise sell you software.
Where there is no evidence we report unassessed, rather than scoring absence as weakness. An honestly named gap is one somebody can close.
“Compliant” should mean there is something to show. Every number we produce is computed from evidence that can be inspected, and every statement the advisor makes cites the evidence it rests on.
Where there is no evidence, we report unassessed — never a failing score. Scoring absence as weakness manufactures findings a client cannot act on, and it destroys trust the first time somebody checks.
The people who depend on the data should not be the last to know. Progress, blockers and live status are visible without anyone having to ask for a status update.
The platform is built, deployed and running. Everything described on this site is working software you can be shown, today, against your own data.
Each statement here maps to something demonstrable in the product. We put our evidence in the demonstration rather than in a logo wall.
Every engagement starts with a defined pilot and success criteria agreed in writing. Our clients shape the roadmap directly.
No fine-tuning on client data, ever.
Oregon AI works in a private environment. Your business data is never exposed to the internet.
Your data is never shared with other organisations or used to answer anyone else.
We work with the data platforms you already have. Nothing is moved or copied.
The fastest route is the pilot request — it reaches us directly and tells us enough to give you a straight answer on fit.
One part of your business, success criteria agreed up front. Request a pilot
Success criteria we agree up front — and you keep the result either way.