Policy : Share AI Knowledge Openly Across the Organisation

Commitment to Open AI Knowledge Sharing AI capability is not built by individuals or isolated teams — it is built by communities of practice where knowledge, tooling, patterns, and hard-won lessons are shared freely. In organisations where AI knowledge is hoarded — where teams reinvent solutions to problems others have already solved, where failure insights are suppressed because they reflect poorly on individuals, where successful approaches are kept proprietary to preserve competitive advantage within the organisation — AI development is slower, more error-prone, and less satisfying for the people doing it. Our commitment is to build the cultural and structural conditions for AI knowledge to flow freely across the organisation.

What This Means Open AI knowledge sharing means creating and maintaining the forums, artefacts, and norms that enable practitioners to share what they have learned — both what worked and what did not. It means celebrating thoughtful experimentation, even when experiments fail, as the engine of organisational learning. It means making the tacit knowledge of experienced AI practitioners accessible to those who are developing their capabilities. And it means ensuring that AI learning is not dependent on individual generosity but is embedded in organisational structures that make sharing the default.

Our commitment to sharing AI knowledge openly is built on:

• Community of Practice – We maintain an active AI community of practice with regular gatherings, structured knowledge-sharing sessions, and a recognised role in the organisation's learning infrastructure. Community participation is encouraged and recognised — not treated as optional overhead on top of delivery work.
• Shared Pattern Libraries – Reusable approaches — data pipeline patterns, evaluation frameworks, model architectures, MLOps configurations, fairness assessment methodologies — are documented in shared pattern libraries maintained by the community. Teams begin new work by looking to the library before building from scratch.
• Failure and Learning Retrospectives – When AI experiments fail, do not deliver expected outcomes, or surface important lessons, those learnings are documented and shared across the community — not suppressed or confined to the immediate team. Blameless post-mortems on AI failures are a standard practice and their outputs are organisational property.
• Experimentation Recognition – Thoughtful AI experimentation is recognised and valued, regardless of outcome. People who try novel approaches, test new techniques, and share what they discover — including negative results — are celebrated for advancing organisational capability. Risk-aversion driven by fear of visible failure is actively countered.
• Cross-Team Collaboration – Teams working on related AI problems are connected to each other and encouraged to collaborate. We prevent the reinvention of solutions to problems already solved elsewhere, and we facilitate the cross-pollination of approaches between teams working in different domains.
• Learning Path Development – We invest in structured learning paths for AI practitioners at different experience levels, drawing on internal expertise as well as external resources. Senior practitioners contribute to learning path content and mentoring as a recognised part of their role.
• External Knowledge Integration – We monitor and integrate relevant external AI advances — research publications, open-source tooling, regulatory developments, industry case studies — into the organisation's knowledge base. External knowledge is curated, contextualised, and made accessible rather than left for individuals to discover independently.

Why This Matters AI knowledge sharing has a compounding return. Every lesson learned and shared becomes a lesson that does not need to be relearned. Every pattern documented and made available becomes a lever that accelerates future work. Organisations that invest in knowledge infrastructure accumulate structural advantages in AI delivery quality and speed that no individual hiring or project investment can match. And practitioners who work in cultures of genuine knowledge sharing — where learning is celebrated and expertise flows freely — report higher satisfaction, greater confidence, and deeper commitment to their work.

Our Expectation Every AI team participates in the organisation's AI community of practice, contributes knowledge artefacts when they develop genuinely reusable insights, and shares the learnings from both successful and unsuccessful AI work. Individual knowledge hoarding is not tolerated; community knowledge building is recognised and rewarded. Sharing AI knowledge openly across the organisation is how we build an AI capability that makes practitioners Happier, more capable, and increasingly effective — together.