Institutional investment professionals were spending weeks to surface actionable insights from alternative asset data. This slowed decision-making, reduced platform adoption, and limited upsell potential. Leadership needed a solution that could accelerate time-to-insight without overloading engineering or confusing users.

I led a deep dive into the end-to-end user journey, conducting interviews with 12 asset managers and mapping how they consumed data.

Key findings
Users struggled to prioritize insights due to volume and noise.
Contextual metrics were scattered across dashboards, making pattern recognition difficult.
Peripheral moments (alerts, comparisons, portfolio anomalies) were largely ignored.

I also assessed technical constraints:
AI models could generate hundreds of insights daily, but the backend limited real-time delivery and dynamic visualization.

Realization

We concluded that the solution needed to:
• Surface the most actionable insights first.
• Template insight delivery so users could scan, filter, and act quickly.
• Integrate seamlessly into existing workflows without requiring training.

I realized that by templatizing insights and introducing an “Insights Feed”, we could cut discovery friction while keeping the system scalable.

Solution

I started by pressure-testing what our Data Intelligence Models could genuinely deliver, grounding every concept in real backend capability to get a clear view of what was possible. Then I mapped customer workflows to AI/ML opportunities that actually mattered.

The focus wasn’t "more data," but smarter moments that change how people decide and act.

I turned those opportunities into prototypes that showed how a single metric could unfold into a narrative worth exploring. That clarity pulled cross-functional partners into alignment and gave leadership a compelling vision of where Insight Explorer could go next.

To enable teams to scale a wide range of insight types, I developed a flexible UI template for the “insight” object. Each insight was tied to a metric, correlated data filters, a value of change the insight could drive, and supporting language or visuals to aid comprehension. To ensure comprehension, I validated various iterations with internal SMEs and customer users.