One governed use case
A single engagement Regent governs end to end — a post-discharge follow-up, a renewal nudge, an overdue-order reminder. Narrow on purpose: a governance layer proves itself on one decision path, not on a migration.
Design Partner Pilot
Prove it on one decision.
A governance layer is not something you evaluate from a deck. The pilot is scoped so you can watch Regent hold a message that should not have gone out, and then read exactly why it held.
UniversalHealthCare.ai is tenant #1 for the controlled trial. First governed traffic is running. Pilot cohort is open.
What the pilot includes
Deliberately narrow.
One use case, one channel, one policy profile. Everything below exists to make a single governed decision path legible and defensible.
One primary channel
WhatsApp on your own Meta business identity, or email. Governed dispatch from your own Meta Business identity.
One policy profile
Your consent definitions, crisis criteria, content policy, quiet-hours window, frequency caps and named escalation target — compiled into the profile the gates actually run.
A Guided Eval first
Before anything connects, we run your scenario against fabricated personas and you read the governed decisions it produces. Every action is simulated. If the semantics do not fit, that is the cheapest place to find out.
Integration support
One push endpoint against your own scheduler. We help you shape the snapshot and validate it before the first real request.
The decision console
The feed, the receipts, the policy view, escalations, suppressions and channel health for your own tenant.
An audit and evidence review
A session with whoever has to defend this internally, reading real governed decisions rather than a slide about them.
The shape
A proposed programme, in four stages.
Stages, not calendar commitments. A 30-day evaluation is the shape we propose; the actual dates are set with the pilot owner against your own readiness.
- Stage 1
Guided Eval
Your scenario, fabricated personas, the full gate chain, every action simulated. No real customer data, no integration and no agreement required.
- Stage 2
Profile and provisioning
Policy profile filled, channel connected on your own identity, templates registered, snapshot schema issued with its validator.
- Stage 3
Governed traffic on one use case
Your scheduler pushes on your cadence. Every decision — sent or not — lands in the console with its receipt.
- Stage 4
Evidence review and decision
Read the audit record with the people who have to sign off, then decide whether it goes wider.
Regent is in build and pilot preparation. Trial go-live for the first cohort is gated on work that is named and tracked — see Trust & data handling for what is built, what must ship before the first live tenant, and what comes after.
What you bring
Four things, and two of them are people.
A pilot that fails usually fails on ownership, not on integration.
A named technical owner
Someone who can push a request and read a response. The integration is small; the ownership is not optional.
A named escalation owner
A real human who receives crisis escalations and can acknowledge them. An unowned inbox is not an escalation target, and Regent will not pretend it is.
Your own WhatsApp Business identity
Your Meta business account and number stay yours. Regent is provisioned into them; it does not resell a number to you.
One decision worth governing
A real outreach your organisation already makes, and a real reason it must not misfire.
Commercial terms
Pricing and cohort terms are set with the pilot owner. There is no public price list for the design-partner cohort, and inventing one here would be the same kind of unsupported claim this product exists to prevent.
- Pre-contract evaluation runs against public website context and fabricated personas. The Guided Eval reads a public website server-side, requires no production integration and no real customer data, and dispatches nothing; any legal or procurement paperwork required by either party is handled separately.
- The Guided Eval costs you nothing, dispatches nothing, and requires no agreement.
- No public booking system is pinned yet — the Guided Eval is arranged by email with the pilot owner.
Before you write the email
Run the eval. It takes a minute and it is on this page.
Pick a scenario, set a crisis flag, move the clock outside the send window. If the outcomes match how your organisation actually thinks about outreach, the pilot conversation is short.