How it works
From scattered to approved.
Cognatum works as one continuous loop. AI gathers and organizes what your company knows, and a named person approves each entry.
Six years, the first two in an accessible place, on non-rewriteable media. Route any exception through Records Management before it is applied. Records Retention Policy 4.2
Records Retention Policy
Version 7 · owner: Records Management
Recognition
Somebody already answered this.
Last quarter. In a document nobody can find. So you answer it again.
The hunt
So you search. Then ask.
SharePoint. Email. Three people who might remember. The person who knew left in March.
The full problem
Companies already possess most of the knowledge their people need to do their jobs. The problem is that they cannot reliably find it, trust it, or reuse it.
Critical information is scattered across documents, shared drives, email, collaboration tools, business systems, and the memories of individual employees. It is often duplicated, outdated, unorganized, unverified, or invisible to the people who need it.
As a result, employees spend countless hours searching for information they know the company already has. When they cannot find it, they recreate work that has already been done, make decisions based on incomplete information, or guess. Sometimes, they simply make something up.
AI has the same problem, only at greater speed and scale. When an AI system cannot find authoritative company knowledge, it may substitute information from the internet, infer an answer from incomplete context, or generate something that sounds plausible but is wrong.
The cost
Then you guess.
Or you rebuild work your company already paid for.
Source A · Retention policy v7
Supervisory records are kept for six years.
Source B · Supervisory procedures v3
Supervisory records are kept for five years.
Conflict held apart · routed to Records Management to resolve, not averaged.
AI, faster
Your AI does the same.
When it cannot find your approved answer, it writes one that sounds right.
Yes. Early withdrawal carries no fee, and no additional disclosure is required for retail clients.
No source · no approver · no date
The stakes
Then someone asks who approved.
In a regulated business, that question has a right answer. Or it has consequences.
What that costs
In financial services, an examiner can ask who approved the wording an AI gave a client. In life sciences, an inspector can ask the same about a medical information response. If the honest answer is nobody, the finding goes on the record. See how this plays out in financial services and life sciences.
The turn
Knowledge needs a home.
The knowledge is not missing. It is scattered, ungoverned, and unusable.
The full solution
Cognatum gives an organization a governed home for what it knows. At its core is a knowledge base where important company knowledge is collected, organized, curated, approved, maintained, permissioned, and made reusable. Cognatum can draw information from the systems where work happens, while preserving a trusted knowledge layer that establishes what the organization actually knows and what is approved for use. That governed knowledge can then be made available wherever it is needed, to employees, business applications, workflows, and AI systems such as ChatGPT, Claude, specialized agents, and other enterprise technologies.
The question
Which retention period applies to client trade communications?
Today, without a governed home
- A slide deck from two years ago says five years.
- A policy PDF on a shared drive says six.
- A chat thread says it depends, and links nothing.
- Nobody knows which one is current.
With one approved entry
Six years, the first two in an accessible place, on non-rewriteable media.
The loop
One loop. Always running.
AI does the work. A person approves it. Nothing goes live without a name on it.
What the eight steps are
The loop has eight steps. AI captures knowledge from the systems where work happens. It structures, cleans, enriches, and improves what it finds. A person approves each entry, and the approval is recorded with a name, a date, and a source. Then Cognatum deploys the entry everywhere it is needed and watches how it performs. Serving runs over MCP, the open standard AI assistants use to reach outside data, and a plain API. The loop maps clause by clause to ISO 30401, the knowledge management standard.
-
01
Capture
Ingest from every source.
-
02
Structure
Normalize the format.
-
03
Clean
Update, deduplicate, reconcile, archive.
-
04
Enrich
Add context and tags.
-
05
Improve
Refine clarity, usability.
-
06
Approve
Human sign-off, audited.
A person holds this step
-
07
Integrate/Deploy
Push live everywhere.
-
08
Reuse
Serve people and AI.
No migration
It organizes itself.
No six-month cleanup project. You curate as you go. The AI handles the rest, in the background.
How that works
You do not stop working to reorganize. You keep answering questions, writing documents, and closing tickets. Cognatum watches those systems and drafts entries in the background. Sorting, tagging, and deduplicating run automatically under human supervision. Your team curates a little each day, and approval stays with a named person. The library builds itself out of work you were already doing. More on the platform page.
Everywhere
Then it goes everywhere.
Your people. Your applications. Your AI. The same approved answer, every time.
Two kinds of reader
Every company now has two kinds of knowledge consumers: people and machines.
What is right? What is current? What is approved? What can I use? What applies to me?
Search alone cannot solve that. AI alone cannot solve that. Governed organizational knowledge does. And once the knowledge is governed centrally, intelligence can happen anywhere.
Read why knowledge needs a governed home and how the connections work.
Cognatum governs the entry
source · version · approver · permissions
Sovrinty
You decide where it lives.
Your hardware. Your country. Your choice of vendors. That is Sovrinty, the control layer on top.
What Sovrinty is
Sovrinty is a separate product from the same company. It works on knowledge that lives in Cognatum: run the knowledge base on your own hardware, move it between countries, and control which vendors connect to it. Cognatum comes first, because control needs something real to hold.
Proof
Every answer shows its work.
Who approved it. When. From what source.
Use disclosure D-214 rev 6. It supersedes rev 5 for all retail communications and carries the current market-risk language. SOP-4471 v3
Common questions
Questions teams ask first.
Do we have to clean up our content before we start?
No. Cognatum is designed to be organized as it is used. You keep working, and it drafts entries from the systems where work happens. The manual steps run in the background under human supervision, and your team curates a little each day.
Who approves an answer?
A person your organization names. Every entry records who approved it, when, and from what source document and version. Nothing is served as approved without that record.
What happens when there is no approved answer?
By default, Cognatum serves approved entries, not guesses. A question with no approved answer becomes work for the loop: the AI drafts an entry from the systems where work happens, and a named person approves it before it is served.
How is this different from enterprise search?
Search points at content wherever it happens to live, so it cannot tell you which version is approved or who approved it. Cognatum stores the governed entry itself. Every answer carries its approval, its date, and its source.
Does Cognatum replace the systems we already use?
No. Work keeps happening where it happens. Cognatum connects to those systems, draws knowledge out of them, and serves the approved version back to people, applications, and AI.
Knowledge governed. Intelligence everywhere.
See it on your own content, in your own environment.