The work.
Request, lifecycle, fulfilment and outcome.
SECURITY & GOVERNANCE
Trust has
an architecture.
PAS — Privacy, Accuracy and Security — is the trust contract. The Trust Layer is how the One Arc Platform implements it.
PThe right information.
The right information.
For the right person.
Tenant, role and purpose boundaries govern access to knowledge, cases and documents. Customer data stays within its authorised context.
- Tenant-scoped data and knowledge
- Role-shaped access
- Purpose-aware information use
AGrounded before answered.
Grounded before answered.
Verified before done.
Answers use approved knowledge and verified records. Missing information is surfaced as a question, escalation or governed knowledge gap.
- Scoped, cited knowledge
- Validation before the decision gate
- Evidence of the required outcome
SPermission to act.
Permission to act.
Evidence of every step.
Identity, authorisation, approval gates and audit controls constrain execution below the user interface.
- Permission-aware system connections
- Consequential actions gated by people
- Separate Case, Approval and Audit Ledgers
THE HUMAN DECISION GATE
AI never
self-approves.
Consequential actions stop at an authorised person. A human approves, rejects or routes the work, and that decision becomes its own evidence record.
The Six Hard Blocks define fixed classes of action that cannot be automated end to end. These boundaries belong to the platform, not to a conversational instruction.
Take the approver’s seat →The authority.
Who made the decision, what they approved and when.
The evidence.
Authority, access and security-relevant events.
01 / The engineering disciplineThe 7 Design Systems
The 7 Design Systems
for AI.
The structural blueprint behind every AI Colleague. Seven disciplines shape how we engineer the product, from scoped knowledge to verifiable execution and human authority.
In the One Arc Platform
The AI Colleague understands context. The Governed Engine controls execution.
02 / The delivery lens
The ADAPT System.
How an AI Colleague is configured for your business: grounded in your policies, scoped to your tools and tuned to your jurisdiction. From the first assessment to measured outcomes.
01 / Assess
Start with the operating reality.
Map the work, review data and knowledge readiness, identify stakeholders and assess risk.
02 / Design
Define a measurable first deployment.
Set the solution architecture, workflow scope, platform configuration, roadmap and success criteria.
03 / Align
Make ownership explicit.
Agree decision makers, stakeholder responsibilities, escalation routes and the communication needed for adoption.
04 / Prepare
Get the team ready to operate.
Prepare approved knowledge, operational procedures, team enablement and training for deployment.
05 / Track
Keep the outcome in view.
Review adoption, operational results and return on investment. Use those findings to guide continuous improvement.
ENTERPRISE REVIEW
Ask the questions
that matter.
How is our environment configured?
Each customer receives an isolated tenant. The Customer Pack defines approved knowledge, policies, roles, workflows, approvals, channels and integrations.
What happens when DINA does not know?
She asks for clarification, escalates or records a governed knowledge gap. Reusable knowledge requires human review and approval.
Can we review deployment and data handling?
Yes. We can walk through the proposed environment, processors, residency requirements and integration scope during your technical evaluation.
What can our security team review?
Discuss tenant isolation, access boundaries, approved knowledge, human decision gates and audit evidence with us as part of your organisation’s technical evaluation.
A COLLEAGUE. A CONVERSATION. A FIRST STEP.