> ## Documentation Index
> Fetch the complete documentation index at: https://docs.integratewise.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Shared Context Principle: One Truth for Humans and AI

> Humans and AI operate on the same governed business context in SpineWorkspace. Same source, same decisions, no translation loss.

The central principle of IntegrateWise is shared context. Humans and AI operate on the same governed business context. Not copies. Not separate vector stores. Not their own private RAG indexes. One shared, governed surface that represents the current, unified state of the business. When the person sees an account at risk, the agent sees the same account at risk. When the agent proposes a next step, it is proposing from the same picture the person is looking at.

## Why Symmetry Matters

Alignment depends on a shared picture. If an agent reasons over a different dataset than the human reviews, every proposal requires reconciliation. The human must translate the agent's view into their own, check for discrepancies, and decide which version to trust. That is not assistance. That is extra work.

Shared context eliminates translation loss. The agent sees the threaded history of events, the attached evidence, the prior decisions, and the current entity state, exactly as the human does. Proposals are grounded in the same reality. Approvals are meaningful because the approver and the actor share the same source.

<CardGroup cols={2}>
  <Card title="Aligned Decisions" icon="arrows-to-circle">
    When everyone, human or AI, works from the same context, proposals and approvals are about the business, not about reconciling different views.
  </Card>

  <Card title="Verifiable Actions" icon="shield-check">
    An agent's action is verifiable because it operates on the same context the human reviewed. There is no hidden state or private reasoning the human cannot inspect.
  </Card>

  <Card title="No Translation Loss" icon="language">
    Intent flows from human to agent without distortion. The agent does not guess what the person meant because it sees what the person saw.
  </Card>

  <Card title="Continuous Memory" icon="infinity">
    Context compounds. Every action, decision, and receipt becomes part of the shared memory. The next human or agent starts richer than the last.
  </Card>
</CardGroup>

## Shared Does Not Mean Identical Access

Shared context is not the same as identical access. Governance controls who and what can see, propose, and execute within the shared surface. Some users see a full account history. Others see a scoped view. Agents may have read access to certain domains and proposal rights in others. The key is that all views derive from the same canonical source, not from separately maintained copies.

<Tip>
  Governance on the shared surface means you can give an agent broad read context without broad write authority. The agent sees enough to be useful, but acts only through capabilities scoped to its role.
</Tip>

## How Spine Fabric Delivers Shared Context

Spine Fabric ingests from every connected system, resolves identities, threads events, and maintains a live operational picture. That picture is the shared context. SpineWorkspace projects it per domain so each user sees what they need. The Spine Agent Network gives agents the same view, not a derivative or excerpt.

The result is one continuity layer, many projections, zero copies.

## Shared Context in Practice

A customer success manager opens an account in SpineWorkspace. They see the full threaded history: the deal, the onboarding, the support tickets, the product usage signals, the renewal conversation. Twin, their personal AI operator, sees the same assembled context. It proposes a prep doc for the upcoming QBR. The CSM reviews the proposal. It was drafted from the same context they are looking at, so review is fast. They approve. Twin delegates to a domain agent that schedules the meeting and updates the record. The receipt joins the shared thread. The next person who opens that account sees the QBR prep, the meeting, and the outcome, all in one place.

This is what shared context enables. Not automation alone. Collaboration between human judgment and AI execution on a single, governed surface.

## Related

* [Business Continuity vs AI Memory](/product/business-continuity-vs-ai-memory)
* [Human + AI Operating Model](/product/human-ai-operating-model)
* [Canonical vs Operational](/product/canonical-vs-operational)
* [SpineWorkspace](/product/spine-workspace)
* [Twin](/product/twin)