Blog
Thinking about enterprise knowledge, AI trust, and the future of how teams find information
Written by the team building InternalWiki. No ghostwriters. No filler. Just what we're learning as we build the enterprise answer layer.
Permission-aware retrieval: why output filtering isn't enough
Most AI tools filter sensitive content from output. We enforce permissions at the database query level. Here's the technical difference and why it matters.
Document age is not the same as staleness
A five-year-old contract is still valid. Yesterday's Slack message might already be wrong. Why we classify freshness by document type, not by date.
What happens when you ask InternalWiki a question
From question to cited answer in under a second. Here's exactly what happens at each stage — no black boxes.
Why InternalWiki exists
Built from direct experience at a property family office where document age was being confused with staleness — and it was costing real money.
Hybrid search: why semantic alone isn't enough
Semantic search understands meaning. Keyword search catches specifics. InternalWiki uses both — here's why and how they combine.
Get new posts delivered
Product updates, engineering deep dives, and what we're learning about enterprise AI. No spam. Unsubscribe anytime.
Join 200+ subscribers