Knowledge Base as a Product (KBaP) is Yaron Helfer’s lens for treating organizational knowledge as a maintained product. Start with the decisions it supports, assign ownership, test usefulness, and repair gaps through a feedback loop.
Begin with the knowledge user
The user may be a customer support specialist, a product manager, or an agent acting within a defined scope. Each needs different information at a different moment. Define the job before deciding how many documents to connect.
Ask what a successful answer helps the user do next. If the answer is readable but cannot support a decision, the knowledge experience still has a problem.
Give knowledge an accountable owner
Separate content ownership from system ownership. A platform team can maintain retrieval while a domain team owns whether a policy is still correct. Make those responsibilities visible.
A useful record includes its source, owner, scope, status, and review trigger. Add dates when they clarify validity. Preserve the distinction between an approved instruction, an experiment, and an unresolved question.
Evaluate usefulness, not document volume
Build a small set of representative questions and realistic tasks. Include ambiguity, conflicting sources, missing answers, and permission boundaries. Use those examples repeatedly as the knowledge changes.
The recommended measures are answer usefulness, source reliability, time to a dependable decision, and correction turnaround. More documents can improve coverage while increasing contradictions; volume alone cannot tell you which happened.
Close the learning loop
Let people flag an answer that is stale, incomplete, or wrong. Route the report to an owner, and make the correction visible to the people who depended on it.
AI needs accurate knowledge like a car needs fuel. The practical implication is ongoing maintenance. Budget for it as part of the product, rather than assuming a one-time cleanup will keep the system useful.
Questions worth asking
Is KBaP a vendor product?
No. It is Yaron Helfer’s product-management lens for dependable organizational knowledge.
How is it different from documenting more?
It starts with a user decision and evaluates whether maintained knowledge helps that decision, rather than measuring document production.
Sources & further reading
These sources inform the context. Recommendations and frameworks are editorial analysis, not claims of proven results. AI-assisted draft prepared for Yaron’s review; personal experience claims require his confirmation.