SharePoint is already paid for. That is the strongest argument for it, and in most conversations the only one nobody says out loud.
If you license Microsoft 365, you have SharePoint. No new procurement, no new contract, no discussion with IT about yet another vendor. Putting knowledge management on SharePoint therefore does not feel like a decision. It feels like common sense. The question is what that common sense costs once you finish the calculation.
What searching costs before anyone finds anything
Employees spend an average of 5.3 hours a week waiting for information someone else holds (Panopto). Not for files someone filed, but for answers that sit in a person's head.
A worked example, to make the order of magnitude tangible. Take 500 knowledge workers, 45 working weeks a year and an internal hourly rate of 60 euros. That is 119,250 hours, or roughly 7.2 million euros a year, tied up in waiting for knowledge. The assumptions are ours, the hourly rate is deliberately low, and you should replace the figures with your own. But even if you take only a tenth of it seriously, that is the amount any solution is up against.
What matters is what sits inside that number. It does not arise because documents are missing. SharePoint environments in mid-sized organisations are full. Overfull, usually. It arises because the information people are looking for is not in any document.
Why SharePoint is still the first choice
This is not a criticism of SharePoint, and a post that pretended otherwise would be dishonest.
SharePoint is good document management. Versioning, permissions, approval workflows, integration with Teams and Outlook: it works, and it works inside an environment your people already use every day. Anyone who decides against it has to explain why a second system is worth the trouble. That burden of proof is fair. Most knowledge management projects launched around SharePoint failed on exactly that question, not on technology.
The three routes usually costed
When organisations notice that their SharePoint is full but answers no questions, there are three typical responses.
The first is a clean-up project. New site structure, metadata, a taxonomy, often with outside support. That costs consulting days once, plus the time of every department re-sorting its content. The result is better filing. It does not answer the question that is in no document.
The second is an AI assistant on the existing content, usually Microsoft Copilot. It finds what is there faster and phrases it fluently. What is not there it cannot find, and it answers anyway. The post on Copilot versus custom enterprise AI covers this in detail.
The third is waiting for the next Microsoft product. Viva Topics was meant to close exactly this gap and has been discontinued since February 2025 (Microsoft). Anyone who built on it is back at the start.
The line items no calculation includes
All three routes get costed against licences or project fees. These items almost never appear:
- Maintenance without owners. A document on SharePoint has an author but no owner who reviews it two years later. After three years nobody knows which of four versions applies.
- Questions to the same few people. When the repository has no answer, people ask the colleague who knows. That moves the cost from searching into the calendars of your most experienced staff, where nobody measures it.
- Knowledge that was never written down. According to Panopto and McKinsey, 42 percent of what people need to know for their job exists in no document at all. No clean-up and no search reaches that share, because it does not exist in the system.
The third item is the largest, and it is structural. SharePoint is built to manage documents someone wrote. Experience is not written. It is told, usually when someone asks, and then it is gone again. How that plays out in individual roles is described in the post on single points of knowledge.
The tipping point is the kind of question
Listen for a week to what gets asked in your organisation. Not searched. Asked.
As long as the most common questions are "Where is the current template?" or "Who owns this form?", SharePoint is the right answer and a good clean-up project is worth the money. Once the expensive questions become "How do we handle this case?" or "Why does site B do it differently?", better filing is no longer enough. What is missing is a layer that captures experience, verifies it and makes it available to people and to AI applications alike.
What is needed alongside SharePoint
That layer does not replace SharePoint. It sits next to it, and ideally your people notice the difference only because they need to call a colleague less often. Omnora Intelligence Management captures knowledge in AI-guided interviews with around 30 minutes of expert time, and each knowledge object carries a named owner, a date until which it is valid, and a record of whether it has been verified. The assistant you already have can reach it through standard interfaces, as a RAG source or via MCP. That does not make it cleverer. It simply gets something verified to work with. In projects we then measure 20 to 40 percent fewer questions to the experts who hold the knowledge; that is our own measurement, not a study. The difference between a document store and a knowledge system is also the subject of the post on the Confluence alternative, because the same applies to Confluence.
So the real cost question is not whether an extra system pays off next to one that is already paid for.
It is who, in your organisation, answers the questions SharePoint cannot answer today. And how many hours a year that person spends doing it.
