Every sport organisation we have worked with shares the same quiet crisis: the knowledge that matters most — the strategy decisions, the player development insights, the lessons from last season's campaign — lives in email threads, shared drives with cryptic folder names, and the heads of people who have since moved on.

When a new analyst joins, they spend their first three weeks asking questions that were already answered somewhere, finding documents that no longer mean anything without context, and recreating work that was done twice already.

We called this the knowledge silo problem, and we built RuNari to solve it.

The Problem Is Structural, Not Personal

The people in these organisations are not disorganised. They work hard. They care about institutional memory. But the tools they use — Slack, email, Google Docs, Notion — are optimised for creating knowledge, not for finding it later.

When knowledge is spread across a dozen platforms, no AI tool can answer "what was our pressing trigger last season" reliably. The search index does not span all your tools. The AI does not have the context. And even if you find the right document, you often do not have the surrounding decisions that made it make sense.

The Insight: AI Needs a Single Source of Truth

Around the time that large language models became useful for knowledge work, we noticed something: the teams that got the most value from AI assistants were the ones who had already centralised their documentation. The AI was only as good as the knowledge you could give it.

That insight drove the first version of RuNari: a Model Context Protocol (MCP) server that indexed your organisation's documents and gave AI assistants a reliable, queryable interface to your knowledge base.

Instead of asking an AI and hoping it remembered the right document, you could connect Claude Code, Cursor, or any MCP-compatible tool directly to your RuNari workspace. It would search your actual documentation — not the public internet, not its training data — and return answers grounded in your organisation's context.

What We Learned in the First Year

The early feedback was clear on two things.

First, document indexing alone was not enough. Organisations needed task management, project tracking, and course authoring — all connected to the same knowledge base. A sport science team did not just want to find their GPS methodology; they wanted to build training courses from it, assign tasks based on it, and track decisions against it.

Second, knowledge silos are a multi-tenant problem. Each club, each federation, each consultancy needed complete isolation. Their knowledge could not leak into another organisation's workspace. That pushed us to build a proper multi-tenant architecture with schema isolation and automated provisioning.

Where We Are Now

RuNari today is the platform we wished existed three years ago: a single workspace where a sport organisation's knowledge lives, is searchable by AI assistants in real time, powers a Courses LMS for staff development, and integrates with the tools teams already use.

We are a small team. We build carefully and ship deliberately. And we are genuinely excited about what sport organisations can do when institutional knowledge stops being the thing that disappears when someone leaves.

If you are building knowledge infrastructure for your organisation, we would love to show you what RuNari can do.