Set up the Company Brain

Admins manage the Company Brain in Platform > Agents > Knowledge. The page lists six areas, each with a one-line summary and a status: Collections, Sources, Brains, Learned facts, Access log and Settings. Selecting an area opens its panel; a brain, source or document opens as a card over the panel. New collection in the page header starts a collection in one step. When something waits on you, a Needs you strip sits above the areas: learned facts waiting for your approval (Review), and sources that are not set up yet, so agents cannot read them (Set up). Who sees the page. Organization admins manage everything on it. A role with view access to the Knowledge area sees every area read-only. The access log, the speed figures and flushing the cache are for organization owners and admins.

​

Brains

Brains lists the Company Brain first, then the specific brains, with what each Contains, its Access (Per person or Shared) and how many Agents use it.

​

Create a specific brain

​

Whose access it uses

• Each person’s own: each person gets only what they can open themselves. Stored documents are filtered by their access, and live sources are read with their own account. Use it for any agent your own people sign in to.
• Shared agent access: anyone using an agent this brain is attached to can read its collections, signed in or not. Use it for public, embedded and community agents whose visitors have no company login. Connected sources are left out, because they need each person’s own sign-in; keep a copy of a folder to make its files available to such a brain.

The Brains table shows Shared in the Access column for a brain with shared agent access, and Per person for the others.

​

Collections

A collection is a set of documents stored in xpander, with an owner and an audience. Documents are parsed into sections with their heading path and page number, and agents cite them down to the page. Collections shows one tab per collection with its documents, their owner, access, version and status.

​

Create a collection

1. Select New collection in the page header.
2. Fill Name and What it holds.
3. Under Who can read it, choose Everyone in the organization, or pick people and groups. Each document can narrow this further.
4. Turn on Let others who can read it add files to let anyone in the audience add documents. Off, only the owner and admins add files.
5. Select Create collection. The Add documents card opens next.

​

Add documents

Add documents on a collection takes three kinds of content:

• Upload: drop files or choose them. PDF, Word, PowerPoint, Excel, Markdown, HTML or images, up to 25 MB each. Scans are read with OCR.
• Paste: a Title and the Text.
• From a URL: the Address of a public page, and an optional Title (the address is used when it is empty). Addresses inside your network are refused.

Each document is processed in the background and shows Waiting, then Processing, then Ready, usually within a minute or two. A document that could not be processed shows Failed with the reason.

​

The document card

Selecting a document opens its card: Owner, Status, Review by, Pages · sections, and Cited in answers this month. A ready document past its review date shows Review overdue. The chunk preview shows how the parser split the document, with the heading path and page each answer cites. Only some people, under Who can read it, limits one document to named people and groups inside the collection’s audience. Off, the document follows the collection. An access change applies on the next question; nothing is parsed again.

​

Versions

New version on the document card adds a new upload, paste or URL as the next version. Only the current version is searched, and the previous one keeps answering until the new one is ready. Each version lists its date, pages and status. Make current (for example Make v3 current) on an older ready version rolls back to it; if it needs re-indexing first, the current version keeps answering until it is ready.

​

Share and delete

Edit collection in a collection’s menu changes its name, what it holds and Who can read it. Delete collection removes it with every document and version; brains that hold it stop finding them.

​

Sources

Sources lists where the brain looks: Collections (built in), the collections stored in xpander, and each live source with what it holds, its Status and Who can read it. A ready source shows Ready and how many people have signed in (for example 188 of 240 people signed in). A source shows Needs connection until it has a connection, and Needs scope while it reads only what you pick and nothing is picked yet. Live sources are read with each person’s own account. Nothing is copied, and a permission removed in the source applies on the next question.

​

Add a source

​

Each person signs in once

There is no shared or service account. Each person connects their own account once, from Settings > Knowledge in Xpander Chat, from the source’s Scope tab when you choose what it reads, or from the Connect button under an answer that skipped the source. Until they do, the brain skips that source for them. The source card’s Connection tab shows your own sign-in, how many people are connected, the timeout and whether any content is stored in xpander.

​

Scope

The Scope tab sets what the brain may read. It still reads only what each person can open.

• All the user can see: everything the person can open. Best for wikis and ticket systems.
• Only what you pick: only what is listed under Include: shared drives, sites, libraries, folders, spaces, projects or single files. Best for file drives, where personal drives add noise.

With nothing picked yet, Include offers a button named for the source, such as Choose shared drives and folders for Google Drive or Choose spaces for Confluence. It opens a picker that browses the source with your own account: tick what agents search instead of typing paths, or add one with Add by path. Once something is picked, Add places opens the picker again, and Leave out places under Exclude ticks what to leave out. The first time, the picker asks you to sign in. Exclude rules apply in both modes: pick Path, Label (for example Confidential), File type (for example xlsx) or Modified date (older than a number of months), enter the value and select Add rule. Let users add their own files (Google Drive, SharePoint and OneDrive) lets each person share single files with the brain through the Google or OneDrive picker. Only the picked files are shared, and exclude rules still apply to them. The scope is applied to the query sent to the source and checked again on every result and every read, so an item outside the scope is never returned, even to someone who can open it.

​

Sensitive sources and cache

The Speed tab shows the source’s p50, p95, timeout rate and cache hit rate, and holds two settings:

• Sensitive: results are never cached, so every answer reads the source fresh. Use it for HR, legal and other restricted material. The Sources table marks it with a Sensitive badge.
• Cache results for: how many seconds results are kept for the same person (0 to 3600, default 60, 0 turns it off). Cached results are per person and never shared, so a permission removed in the source can take up to this long to apply for that person.

Flush cache on the Speed tab clears one source; Flush cache under Speed at the bottom of the Sources panel clears the whole brain cache for the organization.

​

Keep a copy of a folder

For a Google Drive folder, or a SharePoint or OneDrive library folder, the Keep a copy tab copies the folder into a collection. Use it to make those files available to a brain with shared agent access.

1. Enter the Folder (a Google Drive folder link or id; for SharePoint and OneDrive, the library’s Drive id and the Folder item id).
2. Under Copy into, pick a collection or A new collection.
3. Set Who can read the copy. It does not follow who can open the folder: everyone in this audience sees every copied file. Confirm that, then select Keep a copy.

The folder is read with your own sign-in, so you must be connected, and the source’s exclude rules apply to every file. It syncs on a schedule; Sync now starts a sync at once. If your sign-in stops working, the copy shows Needs reconnect until you reconnect or another admin turns it on with their own account. Stop syncing keeps the copy readable by its audience, or deletes it if you choose Delete the copy.

​

Settings

Settings holds what members may do on their own: Whether others may add files to a collection is set per collection, with Let others who can read it add files.

​

Learned facts

An agent can propose a fact, decision or procedure that a conversation established. A proposal is never used in an answer until its owner approves it, and then only for its audience. Learned facts are separate from agent memories, which hold short notes with no approval step. Learned facts shows Waiting for your approval, filtered to Mine to review or Everyone. Each card shows the proposed text, the agent it came from, who can read it and the collection it is saved to.

• Approve adds it as proposed.
• Edit changes the text first, then Save and approve.
• Reject discards it. A rejected fact is never returned.

Who reviews. The owner of the collection the fact goes to; otherwise the owner of the agent’s first attached brain that has one; otherwise the Company Brain’s owner; otherwise an organization admin. The reviewer gets a notification and also sees the fact under Settings > Knowledge > Waiting for your approval. Conflicts. When a proposal contradicts a stored document, the card says This disagrees with that document and the button reads Approve and flag document. Approving asks the document’s owner to update it. Only documents the reviewer can open are named. Every decision is recorded in the audit log with the fact before and after.

​

Access log and speed

Access log lists every brain read: Time, Person, Agent, Source, Document (with version and page) and Result. Search by person, agent or document, and filter by result. Speed, at the bottom of the Sources panel, opens Brain speed, last 7 days: ask time at p50 and p95, the cache hit rate, and the model calls the cache saved.