Knowledge Base
Policies, procedures, and internal guides in one governed library your L&D team owns: organized into spaces and tags, versioned under named owners, and targeted at the people they apply to. Put a policy on a learning path and reading it, confirming it, and taking the training that explains it become one assignment, with a record of who acknowledged which version. Then the assistant answers questions from those same documents, and cites the page it drew the answer from.
The policy, the course that explains it, and the proof it was read — one system.
-
Knowledge Spaces
Policies, procedures, and guides organized into spaces your team names — HR, IT, Safety — with sub-spaces for the detail, like Leave inside HR. Tags cut across them, so Onboarding can gather what HR, IT, and Facilities each own without moving a document. Employees read them in the same app they train in, keep the ones they return to in a saved list, and Arabic documents render as first-class right-to-left pages.
-
Read and Acknowledge
People confirm they have read the policy, it sits on their home screen until they do, and you see who acknowledged which version and when. The acknowledge button appears only on the last page.
-
Versions and Owners
Every entry has a named owner and a review date, so nothing goes quietly stale. A reader who finds a page wrong, unclear, or unreadable reports it from that page; the report names the version and the page, and the owner closes it as resolved or without action. A new version supersedes the old: search and answers return only the current text, prior versions keep their acknowledgements on file, and you decide whether the new one must be read again.
-
Audience Targeting
Visibility follows your org structure: a whole space or a single entry goes to the departments, roles, or groups it applies to. Narrowing a space holds for everything inside it, and what someone browses or searches is only ever what they are cleared to read.
-
Courses from Your Policies
A code of conduct is often better as a course than a PDF — richer than a document, with an assessment built in to check understanding. Build it from the policy already sitting in your library, review the draft, and it feeds the assistant's answers alongside your documents. Publish a revision later and the course is flagged for review rather than quietly rewritten.
-
Reach, Findability, and Gaps
Reporting over 7, 30, or 90 days: how much of the eligible workforce opened a document, whether a search found something, which current documents nobody has opened, and how often the assistant answered. Everything is a count. Small groups are hidden.
-
Policies on the Learning Path
A journey can carry a policy as a step, so the code of conduct, the course about it, and its assessment are one assignment rather than three you coordinate. The path holds at the policy until the reader confirms it from the last page. Reporting says acknowledged or pending — nobody fails a policy read.
-
One Search for Policies and Training
Employees search once and find the policy beside the course, the journey, and the certificate. Results carry only what that person is cleared to read.
-
Reading That Stays in the App
Upload PDF, Word, or PowerPoint. Office files are converted to PDF, so every document reads the same way. Documents open only inside Salalem, with no download or share button, and every page carries a faint stamp of the reader's name, the revision, and the time. A screenshot that leaves the building shows whose screen it came from.
-
The Library in an Employee's AI Client
With the AI assistant add-on, your Super Admin can let employees connect an AI client that supports the Model Context Protocol (MCP), such as Claude or ChatGPT, to the library. Each person signs in with their own Salalem account, and the client can only search and read the documents that person is already cleared to open. It searches by meaning, not only by title. It is off for every organization until the Super Admin turns it on, and turning it off ends every connection already made. What the client reads is then processed by that employee's AI account.
The chain
From the policy to the proof it was understood
-
The policy
Published to the departments, roles, and groups it applies to. This is the link the rest hang from: revise the document and everything downstream follows the version you published.
-
The course that explains it
A policy states the rule; a course shows how it is applied. Author it in the same platform, put it on the same path directly after the policy, and the assistant answers from both.
-
The assessment that verifies it
An acknowledgement proves the document was read. An assessment on the same material shows which parts of it landed, while there is still time to fix the gap.
-
The record
Who acknowledged which version and when, who completed the course, and what they scored — one history per employee, ready the day someone asks for it.
The assistant
Answers from your documents, with the page they came from
Employees ask in their own words and get an answer drawn from the documents you approved, written out as it is generated, with a citation that opens the source at the page it came from. Admins choose which spaces it may read, test the answers before publishing, and see the source list it is answering from.
-
Every answer names its source
The citation beside an answer opens the document at the page the answer came from, so the reader can check the policy itself.
-
It declines rather than guesses
When the approved sources do not cover a question, the assistant says so.
-
It answers within each reader's permissions
A document restricted to one department is only ever cited to that department. The audience you set in the library holds when the assistant answers.
-
Admins decide what it reads
Its sources are the spaces you switch on, the courses you select, and any file you add to it directly — which is filed into the library like everything else, so its citation can be opened. A space is never included because its documents were published — an admin turns it on, and can see what is indexed, stale, or excluded.
Knowledge assistant
Ask about any policy
- Employee
- Can I claim transport home after a late shift?
- Assistant
Yes. When a shift ends after 10pm, transport home is reimbursed up to SAR 120 — attach the receipt to your expense claim within 30 days.
SourceTravel and Expense Policy v4 · page 7
Packaging
How the knowledge base is packaged
-
The library is an add-on
The library, including read-and-acknowledge tracking, is an add-on. Spaces, versions, named owners, audiences, and the record of who read what come with it.
-
The assistant is a second add-on
The assistant, search inside those documents, the connection to employees' AI clients, and the reporting behind them sit on top of the library. Ask us which shape fits your workforce.
Bring One Requirement. See How Its Record Is Built.
Give us an hour and one real requirement — from your rulebook or from a role profile. You will see how it maps to people, deadlines, and evidence on Salalem.
Book a demo