STATE3 for CIOs and IT leadership

A CIO companion that knows your technology estate.

Most of what a CIO gets asked lands between systems. What an application costs sits in one place, what depends on it in another, who owns it in a third, and the contract behind it in a fourth. Answering takes a fortnight and somebody's best recollection, and by the time the answer arrives the question has moved on.

STATE3 holds those connections in one governed model. The companion part (asking in plain language and getting an answer you can take into a meeting) works because of that model underneath it. A language model isn't filling in the gaps.

Book a discovery call See how the AI access works

The standing questions

Things you should not have to commission a report for.

None of these are exotic. The information exists. The answer is hard to get today because it has to be assembled from four systems and three people.

What are we spending, and on what?

Technology spend tracked by contract purpose, with total cost of ownership by vendor and forecast against actual.

Which contracts renew next quarter?

Contracts carried with status, owner and renewal window, so a renewal is a decision you make rather than a date you miss.

Where do we have duplicated capability?

Application-to-capability mapping shows gaps and redundancies, such as two systems that look different until you see they serve the same capability.

Where are we running unsupported technology?

Application and technology lifecycles tracked end to end, with critical and at-risk systems captured rather than inferred.

Which risks have no owner or no review date?

The risk dashboard flags the gaps directly: risks with no owner, no review date and no recorded impact, alongside scored-versus-treated analysis.

What would we affect if we retired this?

Pre-decommission analysis across 14+ entity types, so the consequences are modelled before the call rather than discovered after it.

Which vendors carry the most of our spend?

Total cost of ownership by vendor, with the contracts and the technology in use behind each one, and any risks already attached to them.

Which applications have no clear owner?

Ownership recorded against each application across technology, process and business, and shown as missing where it has never been assigned.

What is changing right now?

Change in flight across the estate, with the workflow, approvers and impact assessment attached to each one.

Why the companion can answer

The experience is the easy half.

Any assistant can be asked a question about your estate. Whether the answer is worth acting on depends entirely on what it is reading from.

Model The organisation, connected rather than filed

Applications, infrastructure, services, capabilities, people, ownership, vendors, contracts, cost, risk and change in one graph.

Five lenses (portfolio, risk, cost, change and service management) over the same model
An answer from one lens is usable in another, because they are not separate datasets
Impact assessment across 14+ entity types, followed past the first hop

Governance The knowledge stays yours

The model is the authority. The assistant is a way of asking it.

Access is scoped to the permissions the user already has
Organisational knowledge stays governed inside STATE3 rather than moving into a model
Queried through the open Model Context Protocol, enabled per tenant

Currency Still true when you ask

A connected model that has gone stale is worse than no model, because people trust it.

Scheduled connectors and ingestion-hub imports staged and reconciled before they're written to the model
A rules engine that applies what it can and queues the rest for a person
Entity-specific imports, direct edits and MCP writes validated as they are written

Not a dashboard, and not a chatbot.

A dashboard answers the questions someone anticipated when they built it. An assistant with no governed model behind it answers anything you ask, which is a different problem. The companion takes your questions, including ones nobody thought to put on a report, and answers them from your own model.

Where this connects