02 — What to build
What your knowledge base app should include
Articles as records with owner, team, status (draft, in review, published, archived) and next review date.
Categories mapped to your processes, not a vendor's default taxonomy.
A review queue showing each owner which articles are overdue.
Role-based visibility
all staff, one department or managers only.
Search and filters by team, category and last-updated date.
Article feedback such as "outdated" or "missing a step", routed to the owner.
A knowledge manager view of ownerless and most-flagged articles.
03 — How to build it
How to build a knowledge base in Henosia
Create a private project
It's a sandbox; nothing goes live until you publish.
Add representative mock data, such as a CSV of titles, owners, teams and review dates. Keep sensitive policies out for now.
Connect a database so articles and review dates are structured data, not loose files.
Describe it in Plan mode
The agent asks clarifying questions, like what happens when an owner leaves, and proposes a plan. Nothing is built until you choose to build.
Build a first version and test it with readers from a few teams, then refine with new prompts.
Publish with access control tied to your company email domain and app-level roles.
An example Plan-mode prompt:
Plan an internal knowledge base for our operations and HR teams.
Each article has a title, body, category, owner, status and next review date.
Only owners and knowledge managers can publish; everyone else can comment or flag content as outdated.
Show each owner their articles past the review date.
HR policy articles are visible to HR and people managers only.Paste it into a private Henosia project in Plan mode. The agent asks questions and proposes a plan first.
04 — Before you go live
Agree on ownership and terms before content moves in
Who owns the knowledge base, and who owns each category? Who edits and who only reads? Who approves publishing the app and later changes?
In a Henosia webinar, Danish e-commerce company 4tal stressed fixing data definitions before automating. The same applies here: agree on what "active customer" or "cost price" means before writing articles that depend on it.
Add an AI knowledge base layer, like suggested answers, once the structure works. Scope any integration per system, data direction and access level; more point-and-click connector interfaces are on Henosia's roadmap. It pairs well with an employee portal.
Start a private project in Henosia, load a sample of your articles, and plan the first version with the people who answer the questions today.
Start a private project05 — FAQ
Frequently asked questions
What is the difference between an internal knowledge base and knowledge base software?
An internal knowledge base is the content: your processes, policies and answers for employees. Knowledge base software is the tool that stores it. Packaged tools give you a fixed structure, while a knowledge base app you build yourself can model your owners, review cycles and visibility rules exactly as your teams work.
Do I need developers to build a knowledge base app?
Not to get started. In Henosia, content owners describe what they need in Plan mode and build a first version without writing code themselves. Bring in IT or a developer when you connect other systems, set up access for sensitive content or prepare the app for wider production use across the company.
How do I keep sensitive articles restricted?
Design visibility in from the start: give each category the roles allowed to read it, then test with users from different teams. Henosia describes access based on your company email domain plus more granular app-level access. Confirm the exact setup with IT before HR, legal or finance content goes into the app.
Can I add an AI chatbot to our knowledge base?
You can plan AI features such as suggested answers or summaries once your content is structured and owned. Answers are only as good as the articles behind them, so keep owners and review dates in place. For critical topics like HR policy, keep a person accountable for what the AI may say.