Systems produce signals. Meaning is harder to find.

Traditional intelligence gathering collects discrete pieces of information and analyses them in isolation. But the most important things happening in any system — a company, a community, a supply chain, a political landscape — are not in the individual data points. They're in the relationships between them. The patterns that give information its significance.

Narrative Intelligence asks a different question: not just what is happening, but how what is happening fits into larger stories about change, power, and possibility. It maps the meaning space — the underlying framework that determines what counts as relevant, how information gets interpreted, and what kinds of responses feel necessary.

"All intelligent systems — human organisations and AI networks alike — operate through narratives. Not stories in the literary sense, but the underlying frameworks of meaning that shape perception and action."

Understanding what we're actually building.

The early phase of this project was almost entirely about asking the right questions. What does it mean to make Narrative Intelligence a product? Who uses it? What's the workflow? What's the minimum form that would still be genuinely valuable?

Discovery workshop — mapping the problem space
Process documentation · whiteboard session · stakeholder mapping

I ran a series of structured discovery sessions to map the system — who the users were, what they were currently doing to get at meaning in complex environments, and where the biggest friction lived. The picture that emerged was clear: people have good instincts about narrative patterns but no systematic way to surface, structure, or share them.

Insight 01
Analysts already think in narrative terms — they just lack tools that match how they actually work. Current tools force them into data formats that strip meaning away.
Insight 02
The most valuable output isn't a dashboard. It's a structured view of how meaning is shifting — what signals are gaining weight, what's being dismissed, and why.
Insight 03
AI and human analysts are not in competition here. They're complementary — AI can surface patterns at scale, humans can contextualise and interpret them.
Insight 04
The platform needs to work across domains — the same meaning structure applies to a policy system, a market, a community, or a technology ecosystem.

From abstract concept to structure.

Turning Narrative Intelligence from a philosophical framework into a product required a lot of iteration on what the core unit of the experience should be. We explored several directions before settling on a model that centres the pattern — not the data point, not the insight, but the relational structure that gives both meaning.

Information architecture — mapping the relational model
IA diagram · entity relationships · user flow
Early wireframes — exploring the core interface patterns
Lo-fi wireframes · concept exploration · 3 directions

The wireframing process moved through three distinct directions. The first was too document-centric — it felt like a CMS. The second was too graph-oriented — impressive but not useful for the primary analyst workflow. The third found the right balance: a structured canvas with a clear reading hierarchy and the ability to drill into relational depth without getting lost.

Mid-fidelity prototype — the pattern canvas
Mid-fi prototype · Figma · interactive testing build

Does it hold up when real analysts use it?

We ran three rounds of usability testing with domain analysts — people who work with complex information environments professionally. The sessions were structured around real tasks: could they take a set of signals and map the narrative structure within 20 minutes? Could they hand off their map to a colleague who hadn't been in the session?

Usability testing session
User research · think-aloud protocol · recorded sessions

The first round revealed a critical navigation issue — users could build a map but couldn't easily navigate within it once it grew beyond a certain complexity threshold. This led to a significant rethink of the spatial model.

By the third round, task completion rates had improved substantially and — more importantly — analysts were starting to use the tool in ways we hadn't anticipated, which is usually the sign that a product concept has found its real shape.

Where we are now.

The platform is in active development. The core pattern canvas is functional, the AI-assisted signal surfacing is in early integration, and the first real-world pilots are being planned. My role continues to span product discovery, development direction, and project management — keeping the strategic thread visible while the build gets more detailed.

Product Discovery IA Design User Research Prototyping AI Integration Usability Testing Project Management Zero to One