AI-Powered Day Trading

—— Scroll Down

AI-Powered Day Trading

—— Scroll Down

AI-Powered Day Trading

—— Scroll Down

—— Overview

A private day trading company wanted to build a next-generation AI trading platform that could analyze markets, generate high-conviction picks, and execute trades automatically without requiring the founders to be glued to a screen. The brief was ambitious: build an intelligent system that runs in the background, surfaces only what matters, and puts the right information in your hand whether you're at your desk or on the golf course. In 12 weeks, we went from no product vision to a live, automated trading system with real signals, real guardrails, personalized AI recommendations, and a mobile-first experience built for life on the go.

Client

Confidential — Fintech / Trading Technology

Date

April 7, 2025

Role

Core Designer

Service

UX Strategy, AI Product Design, Stakeholder Research

Contribution

Responsible for product strategy, UX design, and UI across the fintech trading AI experience, from research through high-fidelity design.

—— Brief

The founders weren't looking to build another trading dashboard to stare at all day. They wanted AI to do the heavy lifting, run in the background, and only interrupt them when it truly mattered.

A 12-week sprint to design and deliver a fully automated AI trading engine, complete with real signals, mobile-first UX, proactive SMS alerts, and the guardrails needed to run live trades with confidence.

60

Reduced signal-to-decision time by 60%

45

Cut pre-trade research from 45 minutes to under 10

8

Weeks from Discovery to Product

Note: The original deliverables from this engagement are confidential and cannot be shared publicly. The screens shown here are reconstructed mocks created to illustrate the product direction and design decisions, not reproductions of the final client work.

—— CHALLENGE

These founders were experienced investors who understood markets but had no shared product vision, no UX, and investors watching closely. The real design challenge wasn't building screens. It was earning trust: making a complex, multi-signal AI feel transparent enough to act on, simple enough to check from a phone mid-round of golf, and safe enough to run with real money.

—— solution

The core insight from research: the founders didn't want to manage a product. They wanted to trust a system. Every design decision was in service of one goal: make the AI invisible when things are going well, and immediate when they're not.

01

Hands-off automated trading

An AI engine that runs continuously in the background, analyzing signals, generating picks, and executing trades automatically with no manual intervention required. The system handles the work. The founders handle their lives.

02

Proactive SMS alerts

Automatic text alerts sent directly to the founders' cell phones when the system needed their attention: an urgent signal, a triggered stop loss, a high-conviction opportunity. No app to open, no dashboard to check. Just a text when it matters.

03

Personalized AI intelligence

Every pick came with a plain-language breakdown of why the system made the call: fundamentals, technicals, sentiment. Once set up, the system also learned each investor's preferences, risk tolerance, and areas of focus, filtering recommendations through the lens of what mattered to that specific investor. Rather than a one-size-fits-all signal feed, the AI tracked market changes relevant to their portfolio and watchlist, surfacing only what was personally significant rather than flooding them with noise.

—— 4 key design decisions

01 Invisibility as the design goal
Early on we had to decide how to structure community within the app. The Facebook Group model was the obvious reference point, but it carries baggage. It feels public, performative, and algorithm-driven in ways that actively work against vulnerability. We landed on Circles. The distinction is intentional: a Circle is intimate, self-selected, and topic-anchored. Each has its own guidelines and keyword filters so members could navigate conversations around specific symptoms or treatment approaches without wading through noise. The name itself matters. A circle implies equal standing, no hierarchy, no algorithm deciding who gets heard.
02 Explainability as the core product, not a feature
Early in the process it became clear that the founders wouldn't trust AI-generated picks they couldn't interrogate. A system making real trades with real money had to show its reasoning. The decision to build plain-language explainability directly into every pick, breaking down the fundamentals, technicals, and sentiment behind each call, wasn't a UX nicety. It was the thing that made the whole system believable. Without it, no amount of good performance would have earned sustained trust.
03 SMS as the primary alert surface
The temptation was to design a rich in-app notification system. We pushed against it. The founders weren't sitting at desks watching dashboards. They were living their lives and needed the system to reach them wherever they were, immediately, without requiring them to open anything. Designing proactive SMS alerts as the primary interruption layer rather than a secondary one was a meaningful departure from the standard app model, and it was the right call for this specific user and this specific use case.
04 Domain immersion before design
Before touching a single wireframe, the process required going deep on trading fundamentals: what signals actually move investors, how risk tolerance varies, what "hands-off" genuinely meant in practice for these founders. That wasn't research for research's sake. It was the prerequisite for making defensible design decisions in a domain where getting it wrong had real financial consequences. The insight that shaped the entire UX came from that immersion: traders don't just need good picks, they need to understand why the AI made them.

—— PROCESS

WEEKS 1–2
Stakeholder discovery
Worked directly with the founders to understand their trading philosophy, daily workflows, and what "hands-off" actually meant to them in practice. Identified the highest-impact design opportunities.
WEEKS 2–3
Domain immersion and strategy
Went deep on trading fundamentals, studying signals, metrics, and what actually moves investors, to build enough fluency to have intelligent conversations with domain experts and make defensible design decisions.
WEEKS 3–5
UX design and flows
Designed the full product experience across dashboard, picks, alerts, performance tracking, risk controls, and automated trading settings.
WEEKS 5–12
Prototyping, testing and PoC
Iterated on designs alongside development, tested explainability flows with stakeholders, and refined the alert logic until the system felt trustworthy enough to run with real money.

The Override

You can't design trust for a domain you don't understand. Before touching a single wireframe, I went deep, studying trading fundamentals, learning what metrics actually move investors, and building enough fluency to challenge assumptions in stakeholder conversations. The insight that shaped the whole UX: traders don't just need good picks, they need to understand why the AI made them. Explainability wasn't a nice-to-have feature. It was the product.