Innovation #1
Two complementary approaches to capturing what AI needs to reason effectively: the official structure and the actual process.
Large Language Models are powerful but stateless—they lack inherent knowledge of your bank's specifics, policies, and data. The real challenge isn't AI capability. It's ensuring AI has access to the right information when needed.
Every organisation runs on two versions of reality. The Official Record is what IT systems capture: forms filled, workflows completed, approvals logged. It's clean, structured, and utterly incomplete.
The Actual Process is how decisions really get made: conversations, context, reasoning, judgment. It's messy, distributed, and almost entirely uncaptured. It lives in people's heads, email threads, and meeting rooms.
We've spent decades building infrastructure for the Official Record. The Actual Process? It barely exists in any system. Now we want AI to help us make decisions—and we've given it nothing to reason from.
Neither approach alone is sufficient. Together they provide the complete context AI needs to reason effectively.
The Map
The traditional perspective: policies, procedures, processes, controls. Essential in regulated industries because you're expected to have documented, auditable frameworks.
A taxonomy provides the human-understandable structure that gives AI guardrails. It defines what entities exist, how they relate, what governance applies.
Explore Risk Taxonomy →How People Navigate the Map
A new perspective: real business processes emerge from actual decisions made through conversations outside core systems. The "real" process that codified systems fail to capture.
When AI agents work through problems, they trace paths through organisational knowledge. Those traces become data—the context graph learns how your organisation actually works.
Explore Context Graph →"The taxonomy is the map; the context graph captures how people actually navigate that map."
Without the taxonomy, you get chaos. Without the context graph, you get the same rigid systems we have today. The combination gives you structure and learning—guardrails and intelligence.
A systematic methodology that leverages both Risk Taxonomy and Context Graph to build and maintain AI agent context.
Gather Information — Far More Than Before
Traditional capture focused on structured data and formal documents. Context Management requires capturing significantly more: meeting transcripts, Slack discussions, email threads, voice notes—the "event clock" that records not just what happened, but why.
This expanded capture feeds the Context Graph. Every conversation, every decision rationale, every informal discussion becomes data that AI can reason from. The institutional memory that usually walks out the door? Now it's captured.
Organize Knowledge — Using the Taxonomy
Raw captured information is noise without structure. The Risk Taxonomy provides the organizing framework: tagging information to policies, mapping discussions to governance forums, linking decisions to the processes they affect.
This is where the two approaches connect. Captured context is curated against the taxonomy—ensuring every piece of information is findable, traceable, and properly governed. The taxonomy provides guardrails; the context graph provides intelligence.
Deliver Insights — Expanding Your Vision
With both taxonomy structure and context graph depth, AI agents can deliver insights that were previously impossible. Not just "what does the policy say?" but "how have similar situations been handled, and why?"
This is where you expand the boundaries of your vision. The context graph becomes a world model—enabling not just retrieval of what happened, but simulation of "what if?" You can reason from precedent, even when the people who set that precedent have long since left.
The institutional memory that usually walks out the door when experienced staff retire? It becomes infrastructure. The reasoning behind past decisions? Available to inform future ones.
This is what enables 10X, 100X amplification. Not AI replacing human judgment, but AI equipped with the accumulated wisdom of how your organisation actually makes decisions—within its actual constraints, culture, and risk appetite.
Over time, context graphs become something more powerful: world models. The organisation can simulate "what if?"—not just retrieve "what was." That's the difference between a chatbot and a trusted risk advisor.
Context provides the foundation. Skills provide capabilities. Patterns orchestrate workflows. Goal Alignment ensures everything connects to strategic outcomes.