← Back to work

Signal

An autonomous AI agent that investigates markets in real time — turning scattered evidence into a scored, source-traceable intelligence report.

Signal

Awaiting investigation

What should Signal investigate?

Signal will map competitors, opportunities, demand shifts, market signals and strategic risks.

Try it live

The real product, running live — give the first investigation a moment to spin up.

signal-whispers-64.lovable.app

Overview

Founders spend months building before they validate demand. Signal is an autonomous agent that investigates a market continuously — pulling evidence from search, social, trend, and launch data — and synthesizes it into a structured, evidence-backed intelligence report. Built in a one-week sprint for the Agentic & Autonomous Systems track at Zuup Hackathon 2026.

Role
Full-Stack Builder
Team
Team of 3 — Manya Mankad (full-stack build), Ananya Rajhans, Aisha Monsar
Timeline
1-week hackathon sprint
1 wk
Build sprint
6
Pipeline stages
4
Live data sources

The problem

Markets now move faster than organizations — a trend can emerge, explode, and disappear before most teams even notice.

Founder behavior

  • Reddit browsing
  • Random customer interviews
  • Competitor stalking
  • Trend hunting
  • Gut-feel decisions

Consequences

  • False positives
  • Confirmation bias
  • Missed opportunities
  • Slow validation cycles
  • Wasted development time

The problem isn't a lack of information. It's turning scattered information into actionable intelligence.

What I built

Signal is an autonomous market-intelligence agent: give it a topic, an industry, and (optionally) known competitors, and it investigates continuously rather than answering a single query. It maps competitors, demand shifts, opportunities, and strategic risks, then returns a scored, evidence-backed report.

As the team's full-stack builder, I designed and built the product end-to-end — the investigation UI, the FastAPI backend, and the multi-stage AI pipeline behind it — working alongside teammates Ananya Rajhans and Aisha Monsar over the one-week sprint.

Opportunity

Automated Debt Management Solutions

Address a major gap with a high-demand feature and strong differentiator.

Score

65/100

Confidence

60/100

Source · Signal Network

Opportunity

Integration with Budgeting Tools

Enhance user experience and data synergy by connecting with current tools.

Score

65/100

Confidence

60/100

Source · Signal Network

Opportunity

Investor-Grade Financial Forecasting

Serve sophisticated users with advanced financial planning capabilities.

Score

65/100

Confidence

60/100

Source · Signal Network

How it works

01

Evidence Collection

Signals gathered from across the internet

02

Signal Extraction

Noise filtered into measurable signals

03

Intelligence Engine

Signals analyzed in depth

04

AI Synthesis

Patterns connected into a narrative

05

Build Recommendation

The case for and against, scored

Technical architecture

Frontend

Lovable + React

REST client, custom components

Backend

FastAPI + Pydantic

Typed request/response models across customer, market, and competitive extraction

CustomerMarketCompetitive

Data sources

Live evidence

DuckDuckGo SearchYouTube Data APIGoogle TrendsProduct Hunt

AI layer

Gemini 2.5 Flash

Executive synthesis · JSON-enforced prompts

OpenRouter — multi-LLM routing

Output

Structured JSON report

Returned to the frontend over REST · deployed on Render

What it produces

Signal doesn't hand back a yes or a no. It builds the case in both directions — the strongest reason to build alongside the biggest risk, a recommended customer and positioning, the most defensible moat, and a confidence score for how well the evidence supports any of it. The decision stays with the founder; what changes is how much they know before making it.

Opening the full brief shows the evidence underneath — the trend, discussion, and competition numbers each score was derived from — so any claim can be audited back to its source.

Signal Verdict

Confidence

88/100

Comprehensive intelligence on pain points, desired outcomes, competitive landscape, and white space provides a clear strategic direction. Lack of explicit market size data introduces some uncertainty.

Top reason to build

Deliver an authentic cookie taste and texture with optimized macro ratios.

Biggest risk

Intense competition from established brands, highlighted by a competition score of 100.

Recommended customer

Fitness-conscious snackers

Recommended positioning

A premium protein cookie delivering authentic, delicious taste and texture, optimized for high protein and low sugar, crafted with all-natural ingredients.

Best moat

Proprietary recipes and natural ingredient formulations for superior taste.

Open full brief →

Market Brief

protein cookie

Why Signal found this

The protein cookie market offers a strong opportunity to innovate, addressing critical consumer pain points regarding taste and ingredients. While competition is high, clear white space exists for a differentiated product.

Evidence

Trend Growth +26%Discussion Volume 114Competition Score 100Growth 26.22

Signal context

Growth

26

Competition

100

Discussion

114

Momentum

13

Signal strength

Score85
Confidence88

What I'd focus on next

The one-week build proved the core loop: evidence in, structured decision out. The natural next layer is distribution-side intelligence — tracking influencer and community activity as an additional signal source, rather than just founder-facing market analysis.

Takeaways

Building the full stack inside a team sprint

Owning the investigation UI, the FastAPI backend, and the AI orchestration layer myself meant every technical decision — how prompts were structured, how concurrency was handled, where evidence got normalized — was also a product decision.

Structured output over prose

Enforcing JSON schemas on every LLM call, rather than parsing free text, was the single decision that made the pipeline reliable enough to demo end-to-end in a week.

Evidence-first, not confidence-first

Every insight had to trace back to a collected source. That constraint — inherited from the product's own positioning — shaped the backend as much as it shaped the UI.

Still here?

If this is the kind of work you want on your team, let's talk.

I'm always up for a conversation about design, emerging tech, or building things that don't exist yet.

or email me at

Currently in Bengaluru (IST) · open to remote-first roles, and to being somewhere else

See more work