Solo founder building a value-investing tool: the honest why and where I am stuck
I am Zaid, a solo founder in Kiel, Germany, and I have been building invest-like.
The itch that started it: every stock tool I tried gave me one of two things. Either a wall of raw metrics that assumed I already knew how to read them, or a single black-box score with no explanation of how it got there. Neither one taught me how a great investor would actually judge the business in front of me. So I built the thing I wanted. invest-like takes any company and runs it through the documented logic of seven value-investing greats (Buffett, Graham, Lynch, Greenblatt, Munger, Fisher and Terry Smith), then tells you in plain English whether it is a strong, partial or weak fit for each one, with the numbers and the reasoning shown. There is also a Boardroom where four of those investors argue the bull and bear case on a ticker, and an Ask Buffett chat grounded in the real Berkshire letters with citations. The part I care most about is honesty. Most tools ask you to trust their score. I publish the methodology and the full backtested track record openly, so anyone can check the work. It is educational, not advice. I am not telling anyone what to buy, I am trying to give regular people the same disciplined framework the professionals use at a price they can actually afford. Stack is Next.js and Supabase, hosted in the EU. It covers 13,000+ tickers refreshed daily, with a free tier and a public API. Right now I am heads-down on distribution, which is honestly harder than the build was. The specific thing I am wrestling with: getting cited by AI assistants like ChatGPT, Perplexity and Claude when people ask for investing tools, since that is quickly becoming a real discovery channel. If you have cracked AI or LLM discovery for a B2C tool, I would love to hear what actually moved the needle for you. And happy to answer anything about the build or the value-investing side.
Pick five buyer questions a value investing user would actually type, run each in a fresh chat, and log whether the answer names you, a competitor, or nobody. Repeat the same five every week so you see drift instead of one lucky screenshot. If you have never checked, we hand out a free five prompt snapshot sheet for exactly that.
The free five prompt snapshot is on my profile product card. Same five buyer questions, you, a competitor, or nobody. https://normbrytande.gumroad.com/l/quwzbl
The AI-discovery angle is probably the right problem to focus on, but I’d be careful treating it like normal SEO.
For tools like invest-like, assistants usually need a very clear reason to cite you: what category you belong to, what makes the methodology defensible, and what query you should be the answer for.
The risk is publishing more content without making the product easy for AI systems to understand as a specific recommendation.
I’d probably think less in terms of “rank for investing tools” and more in terms of owning a narrower citation angle first.
Happy to map the tighter version if useful. The important part is choosing the exact AI-search queries and pages that make invest-like citeable, not just creating more content.