Designing the knowledge layer for a scaling SaaS company.

Leadership had stopped trusting the docs.

The team wasn't short on documentation. Years of material existed across the workspace — pages, wikis, decision threads, process notes.

The problem was that nobody could confidently say which version was current. New hires learned from outdated material. Leaders asked questions in Slack that had already been answered in writing, more than once.

Knowledge without ownership decays.

The audit traced how information actually moved through the company, and the patterns were consistent:

• Documentation had no defined owners, so maintenance was nobody's job
• The same process was documented differently in several places
• Decisions lived in chat threads and meeting notes rather than anywhere searchable
• Structure had evolved team by team, so nothing was predictable company-wide

Architecture first, then implementation.

Phase one was audit and architecture: mapping the information landscape, deciding what actually gets documented, designing the structure, and establishing ownership and maintenance rhythms.

Phase two was implementation — restructuring the workspace, migrating what was worth keeping, and retiring what wasn't. We stayed inside their existing Atlassian tools rather than introducing a new platform, and chose structures that matched how the company already operated.

A knowledge layer leadership can rely on.

The company ended with a documentation system that has clear ownership, a predictable structure across teams, and maintenance rhythms that stop the decay before it starts.

The knowledge is structured and current — which is exactly the context layer AI tools need before they can deliver anything of organisational value.