Launching an AI research platform to 1,500+ early adopters

An AI-driven research platform for investors, taken from discovery to launch — and later acquired by KoinBX.

Role: Product Manager
Dates: Jan 2025 – Dec 2025
Scope: Full product lifecycle

1,500+

Early adopters at launch

0 → 1

Discovery through launch

MCP

Real-time data layer, API-driven

Context

Assetworks AI built an AI-driven research platform for investors, covering market analysis, news and insights. I owned the product lifecycle from discovery to launch. The company was later acquired by KoinBX.

The problem

Two things stood between an AI model and a useful research product. Its outputs were text-heavy, which made them slow to read and act on. And market questions need current data, which a model on its own doesn't have.

What I did

Replaced text-heavy outputs with visual HTML research reports, to improve comprehension and decision speed for traders and investors.
Architected a real-time financial data layer on MCP, pulling market data through API-driven systems to power insights.
Used feedback loops and behavioural analytics to find friction points, and balanced technical feasibility against user experience when prioritising.
Took the platform from discovery through launch to its first 1,500+ early adopters.

Results

Launched to 1,500+ early adopters.
Research delivered as visual reports instead of walls of text.
Insights grounded in real-time market data through the MCP layer.

What I learned

Format is part of the product. The same insight lands differently as a visual report than as a page of text.
For market questions, fresh data mattered as much as the model.