Sources and availability
Understand which business systems can contribute evidence, what a connection means, and how availability is confirmed.
Your Twin starts with evidence from systems your organisation approves.
The first activation uses the smallest set of sources needed for the chosen workflow. Connecting more systems is not automatically better; each source should have a clear purpose, accountable owner, and permitted organisation scope.
Current source examples
Connectors currently used across configured deployments include:
| Area | Examples |
|---|---|
| Accounting and payments | Xero, SunSystems, Stripe, GoCardless |
| Documents and files | Microsoft 365, SharePoint, Google Drive |
| Working documents | Google Docs, Google Sheets, Notion |
| Email and calendars | Gmail, Google Calendar, Outlook |
| Chat and meetings | Slack, Microsoft Teams, Granola |
| Engineering and projects | GitHub, Linear |
This is an availability guide, not a blanket entitlement. The exact connector, account scope, data path, permissions, and refresh behaviour are confirmed for each activation and deployment.
A connector name is not proof of readiness
A connector can exist in the product without being enabled, authorised, or verified for your organisation. Treat it as ready only after the configured account has supplied usable evidence through the intended deployment.
Modern and legacy systems follow the same rule
A source does not need to be new or cloud-based to contribute useful evidence. The Twin is already working with data from Xero and a legacy SunSystems ledger. Both systems remain authoritative while the Twin turns their data into reviewed measures and processes.
For an older application, the implementation team first establishes:
- which records support the chosen workflow;
- how they can be accessed safely and reliably;
- which organisation or business unit they belong to;
- how often the evidence can be refreshed; and
- what limitations reviewers need to understand.
This lets a business make existing data useful for AI without turning the first activation into a wholesale migration or replacement programme.
What the connection states mean
The setup journey distinguishes four important states:
| State | Meaning |
|---|---|
| Needs action | A person must complete sign-in, consent, mapping, or another protected step |
| Checking | Verification or an initial read is still running |
| Ready | The configured source has supplied usable evidence for the intended scope |
| Failed | A reported problem needs to be resolved before the source can be relied on |
Saving settings is not the same as a ready connection. A real permitted read and evidence check must complete.
Questions to answer before connecting a source
Ask:
- Which first outcome needs this source?
- Which account, organisation, folder, project, or ledger is in scope?
- Who can approve access?
- Should the connection begin read-only?
- How current does the evidence need to be for the intended decision?
- Which person can recognise whether the result is complete and correct?
- Are there personal, sensitive, or regulated records that should remain out of scope?
The implementation team records the answers in the activation plan rather than letting an AI assistant infer them.
What happens after evidence arrives
Connected evidence does not become approved business meaning automatically. The Twin can propose definitions, relationships, processes, and measures that the evidence supports. Your reviewers inspect those proposals before anything is published or activated.
Continue with How the Twin works, or plan your first activation if you are choosing the first workflow and sources.