What exactly does Omni do today?
This page is the precise answer. Everything Omni and Valstra offer, in three honest statuses: Available means working in the product today. Supported means delivered as part of an engagement, on your timeline: Valstra delivers through advisory plus the platform, and these are the capabilities the team stands up as an engagement needs them. Not yet means honestly absent. No dates on anything: a capability statement here is a description, not a delivery commitment.
| Capability | Status | What that means |
|---|---|---|
| The connected model | Available | Processes, people, systems, data objects, and documents in one model, generated from documents, transcripts, and wiki pages, and queryable in plain language. |
| Process Recorder | Available | A Chrome extension that captures a process as someone performs it once (clicks, screenshots, typed and voice notes), exported as a single file that Omni turns into a procedure or a process model. Session-based and consented: the person doing the work chooses to record. |
| Future-state scenarios with coverage checks | Available | Change designed as explicit before and after scenarios. Every step, rule, system, and data object in the transition must be accounted for; progress is blocked past an unexplained omission. |
| AI governance and controls evidence | Available | A structured AI governance module whose decisions become documented processes with owners, systems, and controls visible in the same model as the work they govern. |
| Benefits baseline and measurement | Available | Expected benefits attached to each change at design time and tracked on a scoreboard against the baseline set before the work started. |
| Atlassian connectors | Available | Confluence pages, Jira work items, goals and projects. Four connectors on one grant, running today. |
| Connectors beyond Atlassian | Supported | Files (Google Drive, SharePoint and OneDrive), code changes (GitHub), people (Microsoft Entra, Workday), messages (Teams, Slack), records (Salesforce) and control evidence. Each one is built and tested; setting it up against your tenant is engagement work, because it needs an app registration and consent from your side. |
| Estate discovery from your identity provider | Supported | The software your organisation actually runs, read from enterprise app registrations and installs rather than typed into a spreadsheet. |
| MCP (Model Context Protocol) | Available | Your model reachable from the AI tools your teams already run, without a second copy of it to govern. |
| Packs (prebuilt solutions) | Available | Prebuilt starting points for work most organisations share, so a model starts from a structure rather than a blank canvas. Ask which packs cover your scope. |
| Design, build tracking and release | Available | An agreed design carried through development and build tracking to release, with each change traceable back to the model it came from. |
| Audit register and export | Available | Sensitive reads, changes, approvals and every connector grant are recorded in a register you can filter and take away as a file, so an audit is a query rather than a reconstruction. A row records the act and what it touched, never the value itself, and exporting is itself a recorded act. |
| Export and erasure when an engagement ends | Available | Everything belonging to one engagement leaves in a single bundle and is then erased from every store it reached, both computed from one definition so the export and the erasure can never be different lists. Both acts are recorded. Run by our team as part of offboarding rather than a button you press. |
| Point-in-time restore, one engagement at a time | Available | An engagement can be rolled back to an earlier version of itself without moving any other engagement, and the state being replaced is archived first, so the restore has its own undo point. |
| Data residency and sovereignty | Supported | Your model and the material behind it stay in the region you nominate. Australian owned and operated, and the region is settled during scoping. |
| Bring your own key (BYOK) | Supported | Hold your own encryption keys, rotate them on your schedule, and revoke them without raising a support ticket. Set up with your security team during onboarding. |
| Single sign-on | Supported | Sign-in through your identity provider. |
| CSV and Excel export | Supported | The model’s data taken out in the formats your teams already use. |
| API and webhooks | Supported | Programmatic access to the model, and events pushed into your own tooling. |
| Email notifications | Supported | Updates from the model and the engagement portal delivered by email. |
| AI cost management | Supported | Visibility into what is being spent on AI and software, what it supports, and whether it is returning anything, delivered by our advisory team working in the platform. |
| In-workspace multi-user editing | Not yet | Concurrent editing, comments, and an in-workspace permissions model are not yet available. Engagements collaborate through the shared portal today. |
| Seat utilisation telemetry | Not yet | Omni does not measure how much each seat is actually used. The estate itself can be read from your identity provider; what nobody has logged into for six months is a separate question, and not one Omni answers today. |
If a capability you need is not listed, ask. The honest answer might be “not yet”, and you will get it straight rather than discovering it in a proof of concept. For how the AI inside Omni is constrained (what it refuses to invent, and why), read how we constrain the AI.
Common questions
What does "Supported" mean on this page?
Delivered as part of an engagement, on your timeline, rather than sitting pre-built behind a settings toggle. Valstra delivers through advisory plus the platform, so a Supported capability is one the team stands up for you as the engagement needs it. It is a genuine description of how the business delivers, not a hedge.
Does Omni support single sign-on?
Yes, as a Supported capability: sign-in through your identity provider is stood up as part of an engagement rather than self-configured.
Does Omni continuously monitor what staff do on their screens?
No, by design. Capture is session-based and consented: a person chooses to record themselves doing a task once with the Process Recorder. There is no admin-configured, always-on background capture, and none is planned, because the consented model is the point.
Is Omni certified against SOC 2 or ISO standards?
Valstra does not claim any certification it does not hold. We work to recognised frameworks and can discuss your security and compliance requirements directly; the security page describes the current posture honestly.
See it on your own operation
A demo is not a slide deck: we show you a living model of your own business, and you can put any row on this page to the test.
Book a demo