I spent a month building an AI stock analysis product. Launching tomorrow on several platforms. Here's the honest story.

Hey IH 👋

I'm Mannmay. I run a small financial research firm in India called Greenwich Research. We've been producing institutional-grade stock research for 15 years — our data sits on LSEG Workspace, FactSet, LexisNexis. Hedge funds and asset managers use it.

About 3 months ago I started asking a question that wouldn't leave me alone: why does all the good analysis stay behind institutional paywalls?

If you're a PM at Citadel, you have NLP tools scoring every word a CEO says on an earnings call. You know within hours whether management is hedging, whether the CFO's tone diverged from the CEO's, whether the guidance was precise or deliberately vague. If you're a regular person with some AAPL shares — you get a price chart and a Reddit thread. That's it.

So we built StockDashboard.com.

What it does:

It takes the same kind of management voice intelligence we sell to institutions and makes it accessible to anyone. For 50 US stocks right now, you get:

• AI-generated bull/bear cases and company analysis

• Management sentiment scoring from earnings calls

• A scenario conditions engine — "what would need to be true for NVDA to hit $250"

• Buffett-style questions you should be asking about any stock you own

• Interactive signal widgets where you vote your conviction and then see how thousands of other investors read the same stock

That last part is what I'm most excited about. It's not just AI telling you what to think — it's AI analysis combined with crowd intelligence. You see both, then you decide.

What went wrong along the way:

Our original contractor (full-stack dev + his agency) walked away on Day 1 of the engagement. Literally the day they were supposed to start building. I had to restructure the entire team overnight — brought our engineering head (who was managing a different product) onto the backend, hired a new frontend developer in 5 days, and accepted a 2-week delay on a product that was already on a tight timeline.

That experience taught me something I should've known already: for an early-stage product, one person leaving can derail everything. The team configuration you start with is almost never the one you ship with.

Where we are now:

V1 is live. 50 stocks covered. The AI analysis is running. The signal widgets are collecting real investor sentiment. We're launching on several platforms tomorrow (Thursday).

We have zero revenue. Zero paid users. A small early adopter community that's been giving us feedback for the past 2 weeks. No funding — this is bootstrapped out of the parent company's existing research infrastructure.

What I'm trying to figure out:

1. Does the AI analysis actually feel useful, or does it feel like another GPT wrapper? This is the existential question. If the analysis is generic, nothing else matters.

2. Which feature creates the daily return habit? We built a bunch of engagement mechanics (daily challenge, streaks, shareable score cards) but I genuinely don't know which one will stick until real users tell us.

3. Is the "AI + crowd" combination actually compelling, or is it two half-products duct-taped together? I think it works. But I'm biased.
   

   What I'd love from this community:

   Try it. Break it. Tell me what's missing. If you invest in US stocks at all, spend 5 minutes on a stock page and tell me if the analysis taught you something you didn't know. That's the bar I'm trying to clear.

   stockdashboard.com

   I'll be in the comments. Happy to talk about the architecture, the business model or anything else. No question is off-limits.

   Thanks for reading. — Mannmay