Project A
B2B prospecting platform: intent signals, scoring and AI-generated sequences with source verification. Designed, built and shipped solo.
The problem
A sales team spends its mornings hunting for who to contact, and its afternoons writing messages nobody opens. The signals that matter (a role just opened, a funding round, a LinkedIn post) are scattered everywhere, and stale by the time you find them.
The product
Project A brings it all into one place: a filterable prospect database (seniority, industry, company size, location), real-time intent signals, engagement scoring and outreach sequences ready to launch.
AI generation, with receipts
The core of Project A: every outreach sequence is AI-generated, then verified sentence by sentence. Every claim is tied to a real source: a LinkedIn post, a company news item, cited as a footnote. Sentences without a source are flagged, and a click on "Regenerate (grounded)" rewrites the message using only verifiable facts.
Below, a live generation, recorded on the product in production:
The craft
- Design system first. Every component is built and documented in Storybook before entering the app: fast iteration, guaranteed consistency, visible regressions.
- An unapologetically dense interface. For a prospecting tool, information density comes first: sorting, filters and statuses readable at a glance, in light mode and dark.
- Simple continuous deployment, to iterate in production from the first weeks.
What this project demonstrates
A complete product: positioning, design, front end, back end, database, AI integration with guardrails, production deployment. All designed and shipped by one person. That's exactly what I bring to an engagement: taking a subject end to end and shipping something polished, fast.