/01
Structured signal feed
Funding, product, hiring, and go-to-market events across tracked competitors — normalized, categorized, and impact-rated so a scan takes seconds, not an afternoon.
← All projects
Case study — 03
Meridian Lens.
Forty sources of competitor noise, distilled into one feed and a weekly brief the exec team actually reads.
Client
Meridian Labs
Live product
Timeline
7 weeks
Team
2 engineers, 1 designer
Scope
Strategy, Design, Engineering, Launch
Meridian Labs arrived with what their CPO called a fuzzy brief: “we spend hours a week watching competitors and still get surprised.” No spec, no wireframes — just a real cost and a hunch that intelligence could be a product, not a spreadsheet.
We sharpened the brief before writing code. The insight from discovery: the team didn't need more data — they needed fewer, better-framed decisions. Every feature in Lens exists to move a signal from “interesting” to “here is what we do about it.” Seven weeks later it was live.
Signal in, judgment out.
/02
The weekly brief engine
One click composes an executive brief from the live signal set: headline, most-active competitor, capital movements, and recommended counters. Written like an analyst, generated like software.
/03
Impact triage
Every signal carries a high / medium / low impact rating driven by explicit heuristics — so the team debates responses, not relevance.
/04
Watchlists & trend lines
Star the competitors that matter this quarter; six-week signal-volume sparklines show who is accelerating before the headlines do.
From noise to signal.
0+
Sources distilled into one feed
0
Weeks, brief to live product
0
Brief the exec team reads weekly
Sharpen the brief. Then build.
The rigor started before the code. We interviewed the four people who consumed competitor intel and found they each maintained a private, duplicated tracking doc. The product had to replace four workflows with one — or it would become a fifth.
We shipped the signal feed first and let the team live in it for a fortnight before building the brief engine on top. That sequencing surfaced the real categories, the real impact thresholds, and the phrasing executives actually wanted — the brief reads the way it does because the data model was pressure-tested first.
The brief generator is deliberately explainable: every sentence traces to signals visible in the feed. No black box, no hallucinated claims — intelligence the team can defend in a board meeting.
Next.jsReactTypeScriptNode.jsPostgreSQLLangChainAnthropic ClaudeSupabaseVercel
Next step
Sitting on a fuzzy brief?
One call. A concrete next step within 24 hours. No pitch deck.
Start a conversation →No obligation · Intro call · Reply within 24 hours