For the enterprise
Enterprise AI that tells you when it doesn’t know.
A regulated, audited, sovereign estate cannot run AI that is confidently wrong. Cornerstone answers from your organisation’s own cited facts, declines when the answer is not there, and runs inside your boundary.
Select any building — a venture is a campus, one building per component.
Habitat: the whole estate, live, for whoever runs it. Included in Enterprise.
Deployment and sovereignty
Hosted in the UK by default or the jurisdiction you request, in your cloud, on-premises or air-gappable, with a document limit sized to your estate.
Security and compliance
SSO and SCIM, classification, audit and retention controls, and a review your team can approve.
Administration
People, teams, licences and access in one place, with access that follows the source document.
Relationship and support
Named account management, enterprise support with an SLA, and services for rollout.
What changes for you
Four things move.
Risk
The confidently wrong answer stops reaching the work
A rule scoped to one region cannot come back as though it applied everywhere. When the answer is not in your knowledge it can say so, the only one of seven methods we benchmarked that could.
Audit
Have the evidence before the auditor asks
Every answer carries who wrote it, when, and where it applies. When a fact turns out to be wrong, you can find every decision that rested on it instead of reconstructing the trail by hand.
Capacity
Do more inside the same AI budget
Cost per question holds inside a narrow band so finance can forecast it, and the same spend reaches further as the knowledge grows rather than climbing with it.
Scale
Individuals get faster, and so does the organisation
What one person verifies once becomes a fact every team’s AI reads from. The bottleneck moves out of people’s heads and into infrastructure the whole organisation relies on.
Leader and follower
One person directs. Several AIs work. Every output is checked.
Most organisations run AI one chat at a time: one person, one assistant, one task, and then that person reviews everything it produced. Leader and follower is how you get past that. A person leads. AI sessions follow, each taking a ticket and working it in parallel. Validator checks what comes back before the leader relies on it. The person’s job becomes directing and deciding, not re-checking.
The leader
One accountable person
Decides what is worth doing, directs the work, and reviews verdicts rather than raw output.
Directs
Coordinator
Hands out the work, keeps the record
Each piece becomes a ticket with what is asked, what was decided and what it must not touch. What comes back is written to the same record.
Tickets, in parallel
Draft the Q3 board narrative
Working
Follower AI session · From the Q2 decisions on record
Reconcile the supplier ledger
Working
Follower AI session · Against the onboarding rules
Review the EU contract addendum
Returned
Follower AI session · Two clauses flagged, cited
Each follower reads from Cornerstone and writes back to it.
Every output
Validator
Checks delegated work
Every output is checked before the leader relies on it, and the verdict goes on the record. Speed becomes something you can trust.
Verdicts back to the leader
The Coordinator hands out the work
Each follower takes a ticket with what is asked, what was decided and what it must not touch, and reports back to the same record, so parallel sessions do not step on each other.
Validator checks delegated work
Every output from a follower is checked before the leader relies on it, with the verdict on the record. That is what turns speed into outputs you can trust.
The leader decides
One person reviews verdicts, not raw output, and directs the next round. Accountability stays with a human. The throughput does not.
Observed, not projected
This is how we run our own company. One builder ran three commercial codebases in parallel this way, and 1,146 pieces of work across 28 projects were coordinated through the platform. The numbers are counted from our own Coordinator, and we publish them on the company page.
The Coordinator is in every edition. Validator comes with Enterprise.
How organisations get here
One team first. Then the estate.
Enterprise is where a successful team ends up, not where the conversation has to start.
01
It starts with one team, on the trial
A team loads its own documents, gets cited answers from them, and records what it decides. No procurement, no project, and nothing anyone has to approve yet.
02
It widens because it works
More teams come on as they ask for it. What one team records, others can read from, with access that follows the source document. Leadership sees the whole estate in Habitat.
03
Enterprise, when deployment, security or scale asks
Your cloud, on-premises or air-gappable. Single sign-on and SCIM, classification and audit controls, a higher document limit, and named support with an SLA. Scoped to where you run it.
What Enterprise adds
Everything in Scale Up, plus what an estate needs.
Four parts, scoped to your deployment. Habitat is included; the rest is configured to how and where you run it.
Relationship and support
- Named account management, with a success plan and regular reviews
- Enterprise support with an SLA and a named contact
Security and compliance
- SSO, SCIM and advanced role-based access
- Classification, audit and retention controls, configured to your policy
- Security review and compliance support for the infosec and procurement gate
Deployment and scale
- Your cloud, on-premises, or air-gappable
- A raised or removed document limit
- Managed ingestion of large or complex corpora with Foundry
Capability and services
- Habitat, the live view across the estate, included
- Validator, for delegated multi-agent work
- Workbench white-label, under your own name
- Professional services and training
Sovereignty
Your knowledge stays yours.
For a regulated European enterprise this is not a preference, it is the condition of being allowed to use any of this at all. The same product in all three deployments; what changes is whose boundary it sits inside.
Bring your own model
No token markup, and no dependency on ours. Inference stays your contract with your provider.
Encryption and classification always on
Classification is how the access system works. There is no edition without it.
Hosted where you say
In the United Kingdom by default. If your policy needs another jurisdiction, we host there, with backups in the same jurisdiction. The choice is yours.
Air-gappable
Every part of the platform deploys inside your boundary, so nothing needs to leave the building.
Proof
Seven methods. Same documents. Same model.
We wrote the benchmark and we built the product, so it is measured on our own corpus and we say so. Eight tasks is a small sample. What it shows is that the methods fail differently, and that only one of them can say when the answer is not there.
See the full benchmark →96.7%
found the right document, against 85% for the best alternative
1 of 7
the only method tested that could say when the answer was not there
8
the questions it got wrong, out of 49 it could not answer, published rather than hidden
The procurement pack
Everything your review will ask for, before it asks.
We take your security questionnaire and your questions directly, with the measures in writing. The data processing agreement and the sub-processor list are available on request.
FAQs
Can we bring our own implementation partner?+
Yes. The platform is built to be implemented by a partner, and a consultancy you already work with can run the rollout with us behind them.
Do you work from our paper, or yours?+
We start from our own terms and data processing agreement and will review yours. Governing law is England and Wales unless agreed otherwise.
Is there a minimum size for Enterprise?+
No. Enterprise is defined by what you need, your own deployment, identity integration, a larger corpus or named support, not by headcount.
Can we pilot before an agreement?+
Yes. One team on the trial with no agreement at all, or a scoped pilot with the security review running alongside it.
Can Habitat run on its own?+
No. It reads from the work and knowledge already in the platform, which is what keeps it current without anyone maintaining it.
Who will we be talking to?+
The people who built it. There is no sales layer between you and them, and the technical questions get technical answers.
Bring your security team.
The questions they ask are the ones this was built around. Book a call and we will take them directly, with a straight answer on whether it fits, before anyone talks about a contract.