ORIENT / EXPLAINED
Continuous Cognition, Explained.
Keeping an organisation’s understanding alive.
This is the loop.
Something happens → Orient understands it → compares it with current reality and intent → judges whether anything should change → updates the shared state when appropriate → identifies what else may now deserve attention → keeps watching.
That is Continuous Cognition.
01 / SOMETHING HAPPENS
Different inputs. The same cognitive question.
A document arrives. A watched source publishes research. A person or agent produces work. Each is an event that may change the organisation’s understanding.
A document is added
A strategy document carries claims, decisions, assumptions and commitments.
A watched source changes
A new research paper may support, challenge or qualify something already believed.
The organisation produces work
An article, analysis or agent output becomes another event—with its own sources and lineage.
Orient keeps the original source and its provenance: who or what produced it, where it came from, when it happened, and what authority or trust attaches to it. The next step is to understand what that source actually says.
02 / UNDERSTAND THE OBSERVATION
What does ‘Orient understands it’ mean?
It means identifying what is being asserted, what supports it, what remains unresolved, and how those pieces relate. Orient turns the source into candidate observations that can be examined against the organisation’s existing understanding.
Claims and Beliefs
What is the source asserting? Is it reporting a fact, expressing a belief, making an assumption or offering an interpretation? A claim in a document is not automatically an accepted organisational belief.
Evidence and Provenance
What supports each claim? Which passage, result or source does it depend on? Who produced it, when, and with what authority? Derived arguments remain connected to their underlying evidence.
Decisions and Commitments
What was decided, promised or proposed? Who is responsible, what conditions apply, and is there evidence of authorisation? A suggestion must not silently become a decision.
Questions and Uncertainty
What is unresolved, disputed or missing? Which caveats limit the claim? What would need investigation before the organisation could reach a justified conclusion?
Entities, Concepts and Time
Which people, teams, organisations and ideas are involved? What dates, deadlines and dependencies matter? Does the statement describe the past, the current position or a future intention?
Goals and Objectives
What is the source trying to achieve, and which organisational goals might it affect? An inferred priority remains an interpretation until the appropriate authority establishes it as intent.
Then it asks how the observation relates to existing beliefs.
Does it support a belief, contradict it, qualify its scope, repeat what is already known or introduce something new? Orient retrieves the relevant state to assess those relationships. A contradiction is made visible; it is not automatically resolved in favour of the latest document.
The result is a structured, source-linked interpretation—not a declaration that every extracted claim is true. Judgement and authority determine what, if anything, becomes maintained state.
03 / REALITY AND INTENT
What does this mean for us?
A document is an observation. Its significance depends on the organisation it enters. Orient relates new information to both current understanding and authorised intent.
Reality
What do we know, believe, doubt or disagree about? What has been decided, and which evidence supports it?
Intent
What are we trying to achieve? Which goals, principles, commitments and measures apply—with whose authority, and for how long?
FOLLOW ONE EXAMPLE
A new paper. An existing belief. A research goal.
Imagine a paper arguing that more capable AI systems are becoming harder to inspect. Orient relates it to an existing belief about model legibility, and to an authorised goal: support the development of Continuous Cognition. It asks whether the paper supports, challenges or qualifies that belief—and which product or research decisions could be affected. The same paper may deserve little attention in a team pursuing a different objective.
Given what we currently understand, what we are trying to achieve and what just happened, what—if anything—should now change?
04 / JUDGEMENT
A new observation is not an instruction to change your mind.
Orient assesses relevance, evidence, recency, uncertainty and consequence. Preserving a justified belief is as important as revising one.
Preserve
Keep the existing state when the new evidence does not justify moving it.
Support
Attach new evidence that strengthens an existing belief.
Form
Establish new state when the evidence and authority justify it.
Qualify
Refine a belief’s scope, conditions or degree of confidence.
Challenge
Make contradictory evidence and unresolved disagreement visible.
Supersede
Replace outdated state while preserving what came before.
Open a Question
Record what remains unresolved and what needs investigation.
Abstain
Decline to determine a change when evidence or authority is insufficient.
05 / SEE THE STATE CHANGE
One event. A different understanding.
Try a simple launch scenario. Supplier qualification fails. The forecast changes, but the client commitment remains visible. Agents investigate alternatives; people retain authority over any change to the commitment.
Continuous Cognition in Action
Meetings · messages · expertise
Files · plans · shared knowledge
Studies · search · investigation
Markets · events · regulation
Data · events · internal outputs
Analysis · actions · results
OrientOne Live
Model.
Continuously maintained.
Supplier expected to qualify.
is on track.
March promised to the client.
Launch depends on supplier qualification.
People
Confirm launch
readiness.
AI Agents
Monitor supplier
readiness.
06 / GOVERNED CHANGE
AI can propose. Authority determines what is accepted.
A model’s inference does not give it permission to redefine the organisation. The evidence and approval required depend on the kind of change and its consequence.
An inferred priority
“The team appears to be prioritising European sovereignty.” This is an interpretation to examine.
An authorised objective
“Our objective is European sovereignty.” Establishing or changing this intent requires the appropriate human or authorised process.
Routine changes may be accepted under defined rules. Strategic claims can require stronger evidence. Goals, major commitments and important decisions require the relevant authority. Human ratification is recorded and remains open to later challenge; it is not treated as infallible.
07 / THE TRANSITION IS PRESERVED
What changed our mind?
The old belief is not simply overwritten. Orient preserves the transition: what was believed before, the observation and evidence, the interpretation, the judgement, the authority and the resulting state.
State Before → Evidence → Judgement and Authority → State After
Organisational memory becomes a history of cognition—not only a history of files.
08 / CONSEQUENCES AND LINEAGE
What else depends on this?
Return to the research paper. A team member writes an article using its findings, existing organisational beliefs and the research objective. That article is derived work, not an independent confirmation of its own sources.
Source and Interpretation
The paper supplies observations. Orient relates them to existing beliefs and the research goal.
Derived Work
An article or briefing uses those sources and interpretations. Its lineage remains attached.
Reconsideration
If the evidence changes, dependent arguments, decisions and outputs can be flagged for review.
A changed belief may affect a research question, product decision, roadmap assumption, investor narrative, customer proposal or standing agent objective. Orient can identify what may now deserve attention. It does not need to rewrite them all automatically.
A separate check helps detect accumulated drift: bounded reconciliation reconstructs the relevant state from evidence without seeing the maintained answer. Discrepancies return as governed correction proposals.
09 / CONTINUOUS DOES NOT MEAN CONSTANT
Everything important stays eligible for reconsideration.
Continuous Cognition does not mean an AI thinks about every part of the organisation all the time. It means evidence, state or intent can make the relevant parts eligible for attention.
Triage the Flow
Lightweight systems identify potentially relevant events. Much of the state needs no new computation.
Examine What Matters
More capable models can examine ambiguity, contradictions and meaningful consequences.
Spend in Proportion
High-consequence questions can justify stronger models and human review. Cost and state quality must be evaluated together.
This is the computational strategy: selective attention and sufficient intelligence for the task. Its economic advantage is something to measure, not assume. A system that changes its mind on every input creates churn rather than understanding.
10 / WHAT ACCUMULATES
The record. The state. The history of judgement.
As people and agents work, their actions and outcomes become new evidence. The loop closes—and the organisation gains more than an archive.
The Record
Documents, conversations, sources, actions and outputs, with their provenance.
The State
Current beliefs, decisions, questions, uncertainty, goals and commitments.
The History of Judgement
Why state changed or was preserved, who authorised it, and what happened afterwards.
The research ambition is to turn reviewed transitions and observed outcomes into a judgement corpus for specialised organisational intelligence. Accepted decisions are not automatically correct training labels: their quality must be tested against evidence, later outcomes and human review.
General Intelligence → Epistemic Intelligence → Organisational State Intelligence
The document is an observation. The organisation is the model.
Orient turns the continuous stream of documents, conversations, sources, decisions and actions around an organisation into a governed, continuously maintained model of what it understands, what it is trying to achieve, what has changed, and what deserves attention next.