Any AI you like
Integrafy KnowOS · The decision operating system
Ask a question, instead of opening seven tabs.
The knowledge that decides a sale and doesn't fit in an ERP field: PDF price lists, price history, agreed terms and internal procedures. Available in natural language, with the source cited.
What price did we last sell the A-4471-B to Suministros Norte at?
€36.90, on 14 October. The current price list says €38.15: the difference is the 8% agreed on family A.
- Order 4500019
ERP · sales · 14 Oct - 8% discount on family A
Email from the account manager · Mar - 2026 price list
PDF · Jan
Works with your systems and your AI
- PrestaShop
- Shopify
- Sylius
- Medusa
- Google Drive
- OneDrive
- Claude
- ChatGPT
- Gemini
- Microsoft Teams
- Slack
- n8n
Video · 1 minute
KnowOS in one minute. Ask your company's data.
The challenge
The ERP stores the transaction. It doesn't store the reasoning.
Today
- The price doesn't come from the price list: it comes from what that customer was last sold and what was agreed.
- Price lists live in PDF and Excel. A person reads them, not a system.
- A report ends up as an SQL query, a request to IT and a queue.
- When a key person leaves, their know-how leaves with them.
With KnowOS
- You ask in natural language and the answer draws on the ERP, price lists and email.
- Every answer cites its source and the date it was stated.
- Each person sees what their permissions allow, without waiting for IT.
- Criteria are recorded, with their origin and how long they apply.
From source to answer. One map of the business.
Your systems and your documents come in through the same door and become a living model of the business. On top of it, your team, your AI and your agents ask questions, each with its own permissions.
KnowOS building blocksYour sourcesWho asksAgents with their own tokenLiving model of the businessDigital Sage
OntologyThe map of the business
Digital SageAnswers with a source
HypothesesValidated patterns
Manual and FAQBuilt from your criteria
MCP serverFrom your own tools
AlertsCritical stock
AnomaliesOrders and margins
ReplenishmentDraft purchase orders
ReportsScheduled
ERP and CRMOrders, prices, customers
Integrafy DataThe PIM: attributes and statuses
DocumentsPrice lists, manuals, email
PeopleBrowser, Teams, Slack
Your AIChatGPT, Claude, Gemini
AppsREST and GraphQL API
One knowledge base, any AI
All of your company’s knowledge, in whichever AI each person prefers.
With each user’s permissions, in KnowOS or in ChatGPT, Claude, Gemini or your preferred LLM. Everyone sees the same and knows the same, using fewer tokens and with a single place to manage it.
Everyone knows the same
One context for the whole company.
- The same current criteria for everyone
- The same sources, in any model
- No per-person versions of files
Each person sees their part
User permissions travel with the knowledge, in every AI.
What you can’t see in the ERP, you can’t see in ChatGPT.
Fewer tokens
The model only receives what each question needs.
- No uploading files in every chat
- No hand-made summaries of history
- Predictable cost per answer
One place to manage
Sources, permissions and criteria are managed in one place, not in each AI.
- Connect once, it works for every AI
- Add and remove users in one place
- Switch models without redoing anything
- One audit trail for every conversation
Product
It isn't a dashboard. It's the missing reasoning layer.
Ontology
Your business as named objects: customers, orders, price lists and criteria.
The same data, now with meaning. That's why the AI doesn't guess: it reasons.
How it worksConnect
ERP, PIM, price lists, documents and email, read-only.
Nothing is made up
If two sources contradict each other, it doesn't pick one: it flags it.
Agents
They detect, draft proposals and wait for your approval.
Security
EU server, zero retention and inherited permissions.
Nothing to change
It doesn't replace the ERP or the PIM: it makes them answerable.
- Read-only until the criteria are validated
- No pipelines or data team
- Your ERP and your store stay as they are
- The ontology and the corpus are yours and exportable
The magic isn't in the AI model.It's in the map you give it.
Comparison
It isn't KnowOS or your AI. It's your AI with a map or without one.
| Your AI on its own | Your AI + KnowOS | |
|---|---|---|
| Permissions | Inherits those of each connected app, one by one. | One permission engine over the ontology: role, rows, columns and actions per token. |
| Session | Cold: every conversation starts from scratch. | Warm: ontology, criteria and sources always loaded, the same for the whole company. |
| Structured data | Whatever fits in the conversation. | Exact queries over the full history. |
| Reliability | If it can't find the data, it may fill the gap with whatever seems reasonable. | If no criterion is recorded, it answers that there is no record. |
| Control | The knowledge lives with the model provider. | The ontology and the corpus are yours. You switch models without changing anything. |
Explore KnowOS
Each building block, on its own page.
How it works
Capture, ontology, querying and the rule that governs everything.
Agents
Alerts, anomalies, replenishment and reports.
By area
What management, sales, operations and new starters ask.
Developers
MCP server, REST and GraphQL API, and permissions per token.
Security
Your company's data, inside your perimeter.
Use case
What is an agentic PIM? KnowOS and Integrafy Data applied to the catalogue.
What people usually ask us.
What is KnowOS?
Integrafy's decision operating system: the knowledge layer that connects your ERP, your PIM and your documents, turns them into a living model of the business and lets your team, your AI and your agents ask it questions, with the source cited.
Can I use KnowOS with ChatGPT and Claude at the same time?
Yes. Each person uses the AI they prefer, connected over MCP, or KnowOS’s Digital Sage. They all read the same knowledge with that user’s permissions, so everyone sees the same and knows the same. And because the model only receives what each question needs, fewer tokens are used.
How is it different from connecting ChatGPT or Claude to my files?
In where the rules live. With the AI on its own, each person uploads their files and the context is lost when the chat closes. KnowOS keeps the ontology, the current criteria and the permissions always loaded, the same for the whole company, and every answer cites its source and its date.
What happens if there's no recorded criterion for what I ask?
The system answers that there is no record. It never completes an answer with whatever seems reasonable. And when two sources contradict each other, it doesn't pick one: it flags it so a person can resolve it.
Can a sales rep see the margin by asking the AI?
Only if they can see it in the source system. Permissions are inherited from the source and applied per token: role, rows, columns and actions.
Do I have to change my ERP or my PIM?
No. KnowOS makes them answerable, read-only at first. It can be purchased on its own: all it needs is read access to the ERP and the documents.
Show us a question that takes an afternoon today.
A 30-minute session on a real case: a historical price list, an internal procedure, a query that today ends up as a request to IT. You leave with the scope of sources and a fixed price.