Knowledge space
Answers from the company’s documents — for any department, without waiting on colleagues
For support, sales, legal, HR, finance, marketing and management: a knowledge space where questions are answered from your company’s policies, contracts, price lists, reports and FAQ — with the document named. No code needed.
← scroll the diagram →
Objection 01
“We do not write code. Why would we need a tool for developers?”
A fair question: everything in the product started with code. Now departments have their own way in.
Each department asks about its own work — from its own documents
A knowledge space is a department’s workplace with no code sections: only “Ask”, deep research, the knowledge base and connections to documents. Below is what departments come with; each has a use case page with real screens.
Support
response time owed to a customer under the contract and the policy, the refund procedure, what we have already answered about this bug
Sales
what terms clients in a segment got, a comparison of proposal versions, the auto-renewal clause in the standard contract
Legal
a contract’s number and term, which contracts have penalties and at what rates, how the new edition of a policy differs from the old one
HR and newcomers
how to apply for leave, who grants access, what the first week of onboarding includes
Finance
which supplier payments are due this month, what is in a budget line, which closing documents a counterparty needs
Marketing
what the brand book says about the logo, whether a text matches the positioning, a summary of presentations
Management
which decisions were made on a project over the quarter, discrepancies between department reports
Product managers and analysts
what was agreed with the stakeholder, where requirements depart from the policy, research for a release
Objection 02
“Everything is already in Notion, on Drive and in department folders. We do not need another tool”
Nobody is going to move documents, and that is right.
Documents stay where they are — the AI agent reads them in place
A connection is access to the sections and folders your IT specialist has opened: reading, plus comments for answers and new documents for research. Nothing is copied in advance: when a question comes, the AI agent searches for and reads exactly what the answer needs. Ready connections: Notion, Confluence, Google Drive, Figma, folders on a server and network drives, trackers.
Knowledge sources
- Confluence
- Figma
- Files on disk
- Google Drive
- Notion
- Internet
- Your own internal service
Trackers and discussions — a question as a comment, the answer in the same place
What is read
- Word
- Excel
- PowerPoint
- Google Docs
- Google Sheets
- Google Slides
- network folders
- text files
- attachments to a question
14 integrations in the catalogue, installed into a program you already have, with no new software; a system outside the catalogue — a CRM, Bitrix24, 1C, a messenger — your IT specialist connects with your own tool, and it is available to the AI agent at once. Documents are read where they are — nothing is copied in advance. Full catalogue →
Objection 03
“How do I check that an answer is right?”
An answer without a source is an opinion. The team has enough opinions already.
Every fact comes with a link to the document, and discrepancies are named
A support agent asks how quickly to respond to a customer on the “Business” plan. The AI agent reads the support policy, the standard and signed contracts and the FAQ, gives 4 working hours, notices that the FAQ promises 8 and is out of date, and cites every document with its section. What is not in the documents is not in the answer.
← scroll the frame →
Objection 04
“How is this different from a chat with a cloud model where you can paste a document?”
You can paste one document. A department cannot work that way.
How it is now — and with a knowledge space
A chat answers from what was pasted into it and forgets it within the hour. A space answers from the company’s documents, which nobody had to move, remembers the department’s rules and shows where every fact comes from.
| How it is now | With a knowledge space | |
|---|---|---|
| Where to look for the answer | In five folders, three wikis and with the colleague who “definitely knows” | One question — an answer from the company’s documents, with a link to the file and section |
| If documents disagree | Nobody notices until a deadline slips or a customer leaves | The discrepancy is named in the answer: both documents, both values |
| You need a document, not an answer | Half a day compiling a summary from five files by hand | Research: a document with links, checked by a second AI agent, in PDF or straight into Notion |
| A new employee | Asks in chat and gets a retelling from memory | Asks where they read the policy and gets a link to the policy |
| What remains afterwards | Knowledge leaves with the people | The department’s rules, facts and pitfalls — in a knowledge base that both people and the AI agent see |
| Who sets it up | Every employee has their own chat with a cloud model and copy-pastes documents | Your IT specialist connects the documents once; employees simply ask |
Objection 05
“Who will set this up?”
The department has no administrator of its own, and that is fine: there is one technical step, and IT does it.
Three steps, only one of them technical
A space is created by the organisation owner or a member with the “Add repositories” right — that is what the app calls the right to create both projects and spaces. Connecting documents is a one-off setup done by your IT specialist following our instructions. After that the department works on its own.
Create a space — the owner or a member with the “Add repositories” right
One field — a name. The space appears at once; create one per department or one for the whole company.
Connect the documents — your IT specialist, once
A small program is installed in your network; in it, a connection is picked from the catalogue — Notion, Confluence, Google Drive, folders, a tracker — and what to read is marked. A system outside the catalogue is connected by IT with a custom tool — separate work following our documentation.
Write down the rules — any member
The knowledge base holds what must always be taken into account: how to answer, what not to offer, where things are.
← scroll the frame →
Objection 06
“Employees will not go to yet another app”
Nor should they: the question is asked where people already work.
Ask from a document, a ticket or the app
A comment under a Notion or Confluence page, in a Google Doc, under a Figma design or in a tracker ticket that mentions the AI agent is a question, and the answer arrives in the same discussion. Research is ordered the same way: the finished document appears next to the original.
In the app
The “Ask” section of the space: question, answer and follow-ups, like a conversation.
In a comment on a document
Mention the AI agent under a Notion or Confluence page, in a Google Doc or under a Figma design — the answer arrives there too.
In a ticket
Jira, Redmine and other trackers: a question as a comment, an answer as a comment, where support and development already work. CRM and helpdesk systems connect with a custom tool, see below.
Objection 07
“I need a document for the head, not a chat answer”
Nobody wants to compile a summary of five documents by hand.
Deep research builds a document from your sources
Two AI agent roles: the analyst reads the documents and writes; the reviewer checks every statement against the same sources and raises remarks; the analyst corrects. The result is a document with links and the discrepancies named, exported to PDF or as a text file.
← scroll the frame →
Objection 08
“What from our documents goes outside?”
Security will ask this first.
Only what the AI agent read for the answer — and only with your permission
A connection reads only what your IT specialist has opened to it, and only when a question arrives. The details of what goes outside, where it is processed and on which plans documents go to third-party model providers are on the security page.
- Only what is open to the connection is read — and only for a specific answer
- Nothing is copied in advance: documents stay with you, and the AI agent reads them where they are
- The AI agent changes nothing in your documents: the answer arrives as a comment, an ordered document as a new page alongside
- Rights to a space are granted to members separately: in the app, support cannot see the HR space unless given access
- The organisation log shows who asked questions in the app and when
Anyone who can see and comment on a page, document or ticket can ask in a comment, and everyone who can see it will see the answer — even without rights to the space. Such a question stays on the page itself and does not go into the organisation log. So your IT specialist connects to the space only the sections, folders and projects that outsiders — customers, contractors, guests — cannot see or comment on.
Answers for your security teamOther questions
What departments ask
We have no developers. Is this for us?
Yes. A knowledge space needs neither code nor a repository: it has “Ask”, research, the knowledge base and connections to documents. All you need is an IT specialist who installs the connector in your network once and connects the documents.
We use Bitrix24, 1C and our own CRM — they are not in the catalogue
The catalogue of ready connections is a starting set, not a limit. Your IT specialist connects your own system with a custom tool: takes a ready one or writes it following our documentation — that is their work, not ours — after which the system is available to the AI agent at once, just like the catalogue.
Will the AI read everything we keep in Notion and on Drive?
What your IT specialist has opened to it. The AI agent has its own account in Notion or on Drive, and IT decides what that account can access; individual sections and folders can also be marked. If nothing is marked, the AI agent can see everything its account sees, so the IT specialist marks sections explicitly. The AI agent does not go beyond these limits, and reads a document only when it is needed for a specific answer.
What about documents that must not be shown to all employees?
Do not connect them — or create a separate space with a separate connection and grant rights to it only to those allowed. Rights are granted per space, as per project. And do not connect such a space where outsiders — customers, contractors, guests — can see and comment on pages and tickets: any of them can ask in a comment.
What if the AI makes up an answer?
The answer comes with a link to the document and section — it can be checked in a minute. What is not in the connected documents is not in the answer either: the AI agent will say it found nothing. In research, a second AI agent also checks every statement.
How much does it cost?
You pay for the AI agent’s work, in roubles, by the volume read and written; the cost of a conversation is shown in its header, and adding people costs nothing. For reference: department questions in the demo workspace cost from 0.3 to 4 ₽ on the balanced plan, usually 1–2 ₽; research costs more, and spending is visible as the work goes. Plans are on a separate page; a starting balance is provided to try it out.
Will the AI change anything in our documents?
No. Existing documents are not edited. Buff only adds a comment with the answer and, when ordered, a new page or document with research. The only thing the AI agent writes on its own is entries in the space’s knowledge base; you see them and can archive them.
Will it read contract scans too?
Only documents with a text layer: Word, Excel, PowerPoint, PDF with text, wiki pages and spreadsheets. The AI agent will plainly name a scan without a text layer as unreadable — the product has no text recognition.
We have a product with code and departments without it. Is it all one place?
One organisation: next to the spaces there can be development projects where engineers work with code. Members get rights to what they need — support sees its space, a developer sees the project.
The first step is a name
Create a space and ask your first question
Sign up by email and the space is ready at once. Until your IT specialist connects the documents, the AI agent answers from the knowledge base and from files you attach to the question. A starting balance is provided to try it out.
Until documents are connected, attach a file to the question