Optimised and shared memory for both artificial and biological intelligence.

Own your knowledge. Own your destiny.

OIDA records what is approved, who it applies to and what changed. Keep track of the current decisions, with their source, scope and effective dates. Spot contradictions and supersessions, and build your own proprietary intelligence day by day.

The problem

The contract says one thing. The chat says another. A call changed the plan.

To work out how to proceed, you piece it together: ask who was there, check a document, compare versions. Your AI agent can also find an authentic source and act on an outdated decision.

In OIDA, you record decisions, sources and scope. Proposals await approval, changes keep their history, and authorised AI clients can ask what applies now.

What was said. What was decided. What still applies.

This is what losing control of your knowledge looks like.

Making fewer mistakes, and managing your ontology and your company memory to build an AI of your own, means saving money.

47%

of knowledge workers struggle to find the information they need to do their jobs

Gartner, Digital Worker Experience Survey, 2023

11applications

used on average each day by digital workers, up from 6 in 2019

Gartner, Digital Worker Experience Survey, 2023

25%

of teams’ and managers’ working time goes into hunting for answers the company already has

Atlassian, State of Teams 2025: 12,000 workers and 200 executives

More than 6 in 10 times

when the document the AI retrieves holds a wrong figure, the AI repeats it as true: even when it knew the right answer

ClashEval, 2024–2025: six models compared on over 1,200 questions across six fields

50 hours a month

spent reconstructing decisions if ten people spend fifteen minutes a day

Lost time has a cost. Measure what changes for your team.

Find the decision

Ask which rule applies today and read the source alongside the answer.

Your AI assistant

QuestionWhich database do we use for a new service?

Toolresolve_decision_state

OIDAone answer, with its source
In force

PostgreSQL is the approved database for new services.

ADR 024 · scope new-services · since 1 Feb 2026

  • ADR 024 · Choice of databaseNew services use PostgreSQL.primary

You found the document. Now establish what applies.

Retrieval is not enough.

Search and RAG help find relevant sources. A knowledge base stores them. AI memory can recall them. Acting also requires knowing whether a proposal was approved, by whom and for which scope. OIDA records those steps: a recent or relevant source does not become authoritative on its own.

# client-a-renewal · example
  1. Sales9:12I propose €50,000 a year for Client A’s renewal.
  2. Account manager9:36Let’s take the proposal to the next call.
  3. Sales9:41The €50,000 proposal is in the CRM.

Which amount applies to Client A’s renewal?

Retrieved text · example
  1. AThe €50,000 proposal is in the CRM.Team chat · 9:41
  2. BLet’s take the proposal to the next call.Team chat · 9:36
  3. CThe signed renewal is €40,000 a year.Signed contract · 1 September

State unchecked€50,000: that is the amount in the proposal.

The proposal is authentic, but the contract superseded it.

OIDA registry
  1. In forceThe signed renewal is €40,000 a year.Client A · contract §3 · replaces the €50,000 proposal
  2. UnknownIs Client B’s renewal approved too?no approval on record

Recorded state€40,000 for A. No approval is recorded for B.

Authoritative sources, approval and scope matter.

Know what you don’t know.

A shared memory that grows as your company grows, standardising, ordering and labelling information without pause, can aggregate and analyse it across thousands of sources, surfacing the blind spots and hidden risks your company does not even know it has. Remember that LLMs are “just” very good predictors of the next token: to control your destiny, you have to control them.

Sources read
  • 1,284documents
  • 37workspaces
  • 9systems
Your AI assistant

Which risks are we not taking into account in the move of payments to the new provider?

  1. find_conflictsscope payments
  2. Reading the connected sourcescontracts · minutes · chat · registry
  3. Comparing them against the decisions in force3 uncovered risks

Three risks no decision in force covers.

  1. Automatic renewalThe contract with the current provider renews itself on 30 November.Contract FRN-118 · § 9.2
  2. ConflictTwo decisions in force contradict each other on refunds after 30 days.ADR 041 · minutes of 12 May
  3. UnknownNo approved decision on payment data processed outside the EU.no source in the registry · nothing has been inferred

None of these risks is written in a single document: they show up only when every source is read together.

You need an ontology of decisions.

An ontology is a map of the topics and scopes your organisation makes decisions about. OIDA gives you and your AI agents the tools to keep building and improving that organisational ontology, the one that defines how your company really works.

acme-ontology.mdversion 27
  1. Acme S.r.l.
  2. Precision components · Brescia and Timișoara · 240 people
  3. What we call things
  4. “data store” = “database”
  5. “job” = “production order”changed
  6. What we decide about
  7. pricing → discounts, renewals, enterprise
  8. suppliers → contracts, renewals
  9. security → access, personal dataproposed
  10. Who decides over whom
  11. an ADR stands above the meeting minutes
  12. the minutes stand above the chat
  13. the chat counts as a decisionremoved
  14. States of a decision
  15. in force · superseded · scoped exception · in conflict

2 proposals waiting for an administrator · every version stays in history

The solution

A shared memory that grows as your company grows, and builds your own intelligence, for full sovereignty over your data.

One shared vocabulary, versioned like code.

The ontology says which entities, scopes, states and authority rules exist in your organisation. Agents and people read the same version in force: every change is a proposal until an administrator approves it, and earlier versions stay in the history.

Ontology · shared vocabularyone version in force at a time
Writes
  • people, as a proposal
  • agents, as a proposal
  • an administrator approves
Entity types
  • client
  • service
  • supplier
  • contract
Relations
  • decision → applies to → scope
  • decision → supersedes → decision
Decision types
  • technical standard
  • price
  • policy
  • exception
Scopes
  • new-services
  • legacy
  • team-payments
Ontologyversion 27
Synonyms
  • “data store” = “database”
  • “ADR” = “technical decision”
States
  • in force
  • superseded
  • exception
  • in conflict
Authority rules
  • an ADR above the meeting note
  • the meeting note above the chat
Supersession rules
  • on the same scope, the later decision supersedes the earlier one
Reads
  • every connected assistant
  • every person
  • every automation
Every change is a proposal until an administrator approves it. Earlier versions stay in the history, and agents and people always read the same version in force.

Was it decided, or just discussed.

A current rule, an exception, an outdated version, a conflict and a missing decision. Five illustrative examples: choose a question and check its state and sources.

Illustrative examples. Choose a question and compare the answer and sources.

Get a demo

assistant · connected to OIDA

What is the current remote-work policy for the sales team?

  1. resolve_decision_statescope sales team
  2. Reading the sources in the registryHR policy v6, §2 · Management announcement, Feb 2026
  3. State resolvedIn force

The sales team works on site three days a week.

In force · scope sales team · from 1 Mar 2026 · supersedes Two-day rule, HR policy v5

Three days on site per week from 1 March 2026. Source: HR policy v6, §2. The two-day rule is superseded.

Sources
  • HR policy v6, §2From 1 March 2026, sales staff work on site at least three days a week.Primary
  • Management announcement, Feb 2026Evidence

Architecture and sovereignty

Your decisions stay with you, even when you change AI.

Decisions stay in a registry separate from the AI model. In Cloud, data is in the database we run and each organisation has its own workspace. For enterprise integration we talk directly to your IT department and keep your data in your own infrastructure, with your own policy.

Your AI assistantscompatible, via MCP

Your peopleweb app

Who asksyour people and your tools

OIDAthe decision layer
  • explicit rules for decision state
  • traceable sources and evidence
  • recorded authority and scope
  • No copy of your documentsOIDA stores the source text you submit, decisions and evidence. MCP answers return excerpts and references.

  • No trainingOIDA trains no model on the data in the registry. The dataset stays yours: any use starts from a decision of yours.

Your databasewhere the decisions stay

Cloudone workspace per organisation, data kept separate, Frankfurt

Enterprise integrationyour infrastructure and your policy, agreed with your IT department

Your data stays where you decideOriginals stay in their systems; OIDA retains submitted sources.

The model is the tenant. The structure is yours.

Data sovereignty starts with where data is stored and who can access it: we define those boundaries together for the pilot. AI sovereignty starts with a memory separate from the model, queryable through MCP, the protocol that connects assistants to external tools. You can change compatible assistants while keeping the same registry.

95 %or more of respondents consider private and sovereign AI important. 29% concretely prioritise sovereign AI in the near term
NTT DATA, 2026 Global AI Report, nearly 5,000 senior decision-makers across 30+ markets
  • Boundaries. Where the registry lives and who may read it is your call, settled before the pilot starts.

  • Models. Authorised assistants query the same registry, kept separate from models.

  • Permissions. Access by workspace, actions by role. Applicability scope is not a reading permission.

  • Efficiency. Answers carry excerpts, state and references. Use the pilot to assess the context your agent needs.

Accelerated in Turin, with a scientific advisor from MIT.

OIDA is accelerated within Kakashi Venture Accelerator, an AI-native venture studio based in Turin, with senior scientific advisor Pierfrancesco Beneventano, researcher in machine learning theory at Massachusetts Institute of Technology. The method is described in a publication with public evaluation corpora.

We publish what we learn.

All publications

Works with the clients you already have.

Connect a compatible client and authorise workspace access. Your agent can query decisions, sources and states through MCP. Another authorised AI client can consult the same registry.

Connect your AI

  • ChatGPTin developer mode
  • Claudeweb, desktop and mobile
  • Claude Code
  • Cursor
  • CodexCLI and IDE extension
  • GitHub Copilotin VS Code
  • Any MCP client

Frequently asked questions.

How to start, who approves, who can read, and what works today.

Is OIDA an assistant or a chat?

OIDA is a decision registry your AI assistant can query. Keep working in your existing compatible client. You record the relevant sources: OIDA does not automatically listen to calls or conversations.

How is it different from RAG, agent memory or a knowledge graph?

Finding an authentic contract is not enough if a later agreement replaced it. Search, RAG, knowledge bases and AI memory can provide context. OIDA records approvals, source authority, scope, supersessions and conflicts. Explicit rules resolve recorded state, including unknown or insufficient evidence. OIDA can work alongside retrieval.

Who approves a decision?

By default, owners and admins approve. Member contributions stay proposed. An owner can enable peer governance, allowing members to approve decisions too. Inferences from prose stay proposed. Ontology changes always require separate approval by an owner or admin.

Who can read our data, and where is it stored?

OIDA Cloud stores data in Frankfurt. Authorised members can read their workspace registry. Roles govern actions. A decision’s scope says where it applies, not who can read it. An HR draft restricted to leadership must stay outside a workspace shared with the wider team: per-record confidentiality is not available today.

Does OIDA train models on our data?

OIDA Cloud stores decisions, evidence and history without training models. OIDA Intelligence is a separate pipeline, not connected to Cloud and off by default. Using it requires an owner-admitted dataset, recorded rules and evaluation before a candidate model is approved.

What is “MCP”, and do I need to understand it?

No. It is the standard connection an AI assistant uses to reach an external source. For you, it means giving your assistant OIDA’s address and signing in once. From then on, it can ask which decision applies. Instructions for each client are in the installation guide.

Which clients work today?

Claude Code, Cursor and Codex have completed a real end-to-end run. ChatGPT in developer mode, the Claude apps and GitHub Copilot in VS Code follow the vendor’s published setup path. Any client supporting Streamable HTTP and OAuth 2.1 can connect. Clients that send only a static header cannot.

How much does it cost?

OIDA Cloud is in pilot. Request a demo to agree on the use case, access and pilot terms.

Start with one decision. Check what your AI answers.

Choose a decision that changed recently: an offer, a procedure or a policy. Bring the earlier source and the approved one.

  1. 01

    Create your workspace

    Sign up with your email and organisation name. Confirm your email when requested and sign in to your workspace.

  2. 02

    Connect your assistant

    Follow the installation guide: copy OIDA’s address into a compatible assistant’s settings and authorise access.

  3. 03

    Record, approve, check

    Ask your assistant to propose a map of topics and scopes, then approve it. Record the decision with its source, approve it and ask what applies now.