Lookout

Inspiration

There is too much financial news. Every day, people see headlines about inflation, interest rates, company earnings, oil prices, and breaking news. But if you own only a few stocks, most of that information may not actually matter to you. We wanted to answer one simple question: “What actually moves the stocks I own?” That became Lookout AI, a personalized market radar for your portfolio.

What it does

Lookout AI is a personalized market intelligence dashboard for your portfolio.

Users enter the stocks they own, and Lookout builds a historical “fingerprint” for each stock by studying how it reacted to different types of events. It then shows:

which events have historically moved each stock the most,
how large those reactions typically were,
the historical events behind those estimates,
current signals and alerts relevant to the portfolio,
and contextual explanations that make the data easier to understand.

How we built it

The user first enters their portfolio.

Your stocks ↓ Lookout searches past events ↓ Measures how each stock reacted ↓ Builds a unique pattern for every stock ↓ Shows what matters most to you

We built the frontend using React and the backend using Python and Flask. For every major past event, Lookout checks how a stock moved around that event. It then combines many past examples to estimate how strongly that stock usually reacts. We also use AI to turn these numbers into simple explanations. Importantly, the AI does not make up the financial numbers.

Real market data ↓ Our calculations ↓ AI explains the result

This lets us combine trustworthy calculations with easy-to-understand AI explanations.

Challenges we ran into

One of our biggest challenges was making Lookout truly personalized, so every page changes based on the stocks a user actually owns. We also had to handle different types of market events in one consistent system and make sure we did not show misleading patterns when there was not enough historical data. Finally, we spent a lot of time simplifying the experience so users can understand complex market behavior without needing a finance background.

Accomplishments that we're proud of

We started building Lookout much later than we originally planned, so one of the things we are most proud of is how quickly we were able to turn the idea into a working end-to-end product. In a short amount of time, we connected a personalized portfolio, historical market data, backend analytics, an interactive React frontend, and AI-generated explanations into one experience. We are also proud that Lookout is not just a static demo. The analysis actually changes based on the stocks a user owns, and every result can be traced back to real historical examples. Most importantly, we were able to take something fairly technical and make it feel simple: enter your portfolio, see what has historically moved your stocks, and click any result to understand why.

What we learned

We learned that personalization in financial applications is not just a UI feature, it changes the analytical problem itself. A useful alert depends not only on whether an event is important globally, but on whether it has historically mattered for that particular portfolio. We also learned how important statistical context is when presenting historical patterns. A large average move based on three observations should not be treated the same way as a similar result observed across dozens of events. From an engineering perspective, we learned the value of keeping the frontend relatively thin and letting the backend remain the source of truth for authentication, portfolio state, historical analysis, and event evidence. Finally, we learned that AI is most useful here when paired with deterministic systems. Rather than asking a language model to guess what might move a stock, we can calculate the evidence first and use AI to help users understand it.

What's next for Lookout AI

Next, we want to make Lookout more real-time and more adaptive.

We plan to expand the event library, incorporate richer news and sentiment signals, and automatically update fingerprints as new market events occur. We also want to improve portfolio-level analysis so Lookout can identify when several holdings are exposed to the same underlying catalyst. Another direction is smarter alerting: instead of notifying users about every market event, Lookout could rank alerts based on the historical sensitivity of their specific holdings. Longer term, we envision Lookout becoming a personalized market copilot that continuously learns what matters to your portfolio, explains why it matters, and helps separate meaningful signals from everyday market noise.