Cognitive infrastructure

From information to understanding. From interaction to continuity.

Quuipu is an architecture that organizes the digital experience around intent, context, memory, knowledge, evidence, and decisions, rather than around individual applications.

The cognitive chain

every node stays anchored

  1. 1Signals
  2. 2Context
  3. 3Knowledge
  4. 4Evidence
  5. 5Memory
  6. 6Understanding
  7. 7Decision
“Software stores data. Quuipu preserves context.”Navigation is optional. Intent is essential.
01The problem

Digital
without continuity.

For forty years, software has been organized into containers: files, databases, closed applications. Whether researching, closing a deal, or checking compliance, people jump between dozens of tabs and keep the context in their heads.

Application-centric

Fragmented
  • Organized around applicationsYou jump from CRM to ERP, from BI to email: different interfaces, different logins.
  • Navigation is mandatoryMenu → folder → section → record → edit.
  • Every session resets the contextClose the window and the reasoning is gone.
  • Isolated dataRows and files with no link to the evidence and rules that produced them.

People become the “manual integration” between systems.

Intent-centric · Quuipu

Continuous
  • Organized around intentThe interface assembles itself around the user’s goal and constraints.
  • Navigation is optionalIntent → context → action: what you need surfaces without searching for it.
  • Continuity over timeHypotheses, intermediate steps, and collaborators remain available.
  • Relationships, provenance, evidenceEvery element stays linked to its origin and to the rules that govern it.

Every context switch becomes a persistent element, available in later sessions.

02Cognitive core

Eight primitives.
One graph.

Instead of raw values or unsourced text, Quuipu organizes the experience around eight core elements. Together they form a cognitive graph that connects human intent to machine intelligence.

01

Intent

The purpose

Intent comes before the interface: the path is derived from the purpose.

02

Context

The situation

No action out of context: actors, time, objects, and constraints always linked.

03

Knowledge

Connected understanding

Relationships between sources, not flat rows; contradictions made visible.

04

Evidence

Verifiable anchoring

Answer → explanation → evidence → source document.

05

Memory

Continuity over time

No session starts from scratch unless explicitly erased.

06

Decision

Traceable outcome

Every decision retains its rationale, alternatives, and residual risks.

07

Provenance

Complete history

Origin, transformations, and handoffs in an append-only ledger.

08

Governance

Boundaries and sovereignty

Models propose; human authority retains veto power over execution.

03Architecture

Seven layers
on top of what already exists.

Quuipu operates as a separate layer on top of existing databases and message queues: it turns transactions into living cognitive elements without constraining schemas.

Traditional stack · information

  1. Isolated applications

    context is lost between one system and the next

  2. Relational databases

    overwrites erase time

  3. Raw data and files

    no native causality

Quuipu model · cognitive infrastructure

  1. 07ExperienceIntent-generated workspaces, not static pages.
  2. 06IntentGoals translated into execution paths.
  3. 05ContextState, collaborators, and boundaries in real time.
  4. 04CognitiveKnowledge, memory, and evidence, with their history.
  5. 03RepresentationReadable graphs, timelines, dossiers, and matrices.
  6. 02Governance and trustProvenance ledger, permissions, human oversight.
  7. 01Data and systemsExisting systems read without altering their schemas.

Program D · Enterprise cognitive architecture

Formal specifications for preserving state, traversing temporal graphs, and capturing events non-invasively.

Program E · Digital experience meta-model (EDEM)

The model that defines intent-driven workspaces in place of static pages.

04Cognitive workspace

Not a page.
A space that assembles itself.

When someone expresses an intent, Quuipu generates a tailored workspace that evolves and leaves a trail. Follow the six phases.

Workspace · Regulatory impact analysisPhase 1/6

Stated intent

“Analyze the new regulations and assess their impact on our operations.”

RegulationImpact analysisOperationsCurrent quarter

Assembled for you

  • Relevant documents
  • People involved
  • Regulatory context

Activity

  • New regulation published · source added
  • Impact assessment updated
  • Review with stakeholders
  • Decision recorded

Perspectives

ManagementOperationsLegalRisk
Regulatory
Operational
Financial

Milestones

  1. Regulation published
  2. Management review
  3. Decision

Workspace dossier · signed

Complete history, evidence, and decisions in a single verifiable document

Verified

Illustrative interface

Time-aware knowledge

Facts change: every node records when it was created, when it was observed, when it became valid, and what superseded it.

created_atobserved_atvalid_fromvalid_untilsuperseded_by

Identity and actor kept separate

The person, their account, their formal role, and the role they play in a specific operation remain distinct.

IdentityAccountRoleActor
05Trust and governance

AI is a capability.
Not the architecture.

AI models are probabilistic and change over time. Quuipu keeps them separate from identity, memory, evidence, and governance: models can be swapped out, while the infrastructure stays stable and explainable.

Human oversight

The system prepares the options; people authorize the outcomes.

Explainability

Every suggestion shows why it surfaced and on what assumptions.

Privacy by design

Strict separation between personal, departmental, and company-wide context.

Verifiability

Wherever compliance demands proof, every layer provides it.

Evidence first

Unsourced claims are discarded.

Explicit governance

Processing runs within rule corridors aligned with the EU AI Act and ISO/IEC 42001.

06Ecosystem and research

The experience layer
in the Cosen Lab ecosystem.

Quuipu orchestrates experience and context. It works alongside IRIS Core AI, the governance engine, and IRIS Discovery, the organizational intelligence layer.

Research direction · Quuipu browser

The web, with memory.

Addresses, tabs, and bookmarks haven’t changed in thirty years. The idea is to browse by purpose: to reconstruct not only where you went, but why you set out and what led to the decision.

  1. 01

    Reasoning path

    The chain of articles, sources, and archives consulted stays connected.

  2. 02

    Discarded hypotheses

    The alternatives considered and why they were ruled out.

  3. 03

    Evidence summary

    Citations and sources collected in a verifiable dossier.

Quuipu is a research and development program: some of the features described are still being specified.

Connect what you know, what you do and why it matters.

Request the confidential technical documentation, or explore the feasibility of a pilot project with us.