SOPs in Confluence: why the wiki becomes the graveyard

What the wiki is genuinely good at

Knowledge that reads like an article: decisions and their context, project notes, how-to guides, the why behind the what. Every company needs that shelf, and Confluence or Notion is the right place for it. FLOW does not replace your wiki and is not trying to.

Why processes rot there

A process is structured: steps with owners, handoffs between roles, decisions with conditions, documents attached to specific moments. Flatten that into paragraphs and headings and you lose everything that made it operational. Nobody can filter to their own steps. Nothing flags the page when its review date passes. Two teams describe the same handoff on two pages that quietly contradict each other. And no system can execute or even query a process that exists as prose, which is why your wiki is invisible to every automation and every agent you run.

The result is familiar: the page is written during the documentation push, viewed for a month, and then the operation drifts while the page stands still. Run the graveyard audit on your own space if you want the number.

The division of labor

Keep the wiki for knowledge. Move the processes to a system that knows they are processes: steps, owners, scenarios, sign-off, versions. FLOW imports Confluence exports directly, pulls the steps and roles out of the prose with AI, and gives you a draft living process to approve. The SOP document does not die; FLOW regenerates it from the process, so the readable version and the true version are finally the same thing.

The compounding difference

A year in, the wiki approach gives you more pages. The system-of-record approach gives you a governed portfolio: what is live, what is overdue, who owns what, queryable by your team and your agents over API and MCP. One of these compounds. The product page shows how the import works.